email icon Email this citation


Focal Points and International Financial Markets:
The Maastricht Convergence Criteria

Layna Mosley

International Studies Association

March 18-21, 1998

The author thanks Robert Keohane, Claire Kramer, and Peter Lange for comments on earlier drafts of this paper. Support for the interview research described in the paper was provided by the Social Science Research Council's Western Europe Dissertation Fellowship Program. Comments are welcome (email: mam3@acpub.duke.edu).

A central research problem in comparative and international political economy concerns the implications of economic globalization - and more specifically, of international capital mobility - for national economic policy choice. A large body of recent literature suggests that governments are, at least to some extent, constrained by relatively high levels of international capital mobility (Garrett, 1998; O'Brien, 1992). At the very least, the asset allocation decisions of financial market participants affect interest rate levels, and, therefore, the cost of borrowing for governments and private actors.

Much of the "globalization and government policy" literature - and particularly items in the popular press - suggests that international financial market constraints are quite wide and strong (e.g. Kurzer, 1993), so that governments possess little policy-making autonomy. Others, however, maintain that, despite the strength of financial market pressures, governments maintain policy-making autonomy, particularly in "micro side" areas. Therefore, in a few policy areas, we observe cross-national, "downward" convergence, while in many other policy areas, we observe divergence (Berger and Dore, 1996; Garrett, 1998; Garrett, 1996; Garrett and Lange, 1991; Swank, 1997). The implication of these studies is that the financial market constraint on policy choice in advanced industrial democracies is strong, but narrow - financial market participants who deal with OECD nations consider only a few aspects of government policy when making cross-national asset allocations (Mosley, 1998).

Furthermore, within this "strong but narrow" financial market constraint, governments are able to influence the precise nature of the financial market constraint. One means by which governments do so is through the selection or suggestion of market focal points. By virtue of their actions, governments suggest to markets which economic policy criteria are important; market actors then rely on these criteria when making asset allocation decisions. Hence, although governments are subject to "market pressures," they play a role in determining the nature of these pressures.

In this paper, I examine the influence of governments on financial market behavior with respect to European Union (EU) nations and the Economic and Monetary Union (EMU) convergence criteria. I argue that market participants' use of the Maastricht fiscal convergence criteria suggests one means by which governments influence financial markets -- by developing the criteria against which they will be judged. Therefore, the development of the convergence criteria should be viewed not as a technical process, but as a political process, in which governments made decisions regarding the nature of future financial market constraints. I begin with an examination of the general use of policy indicators by international financial market participants. I then discuss the specific use of the Maastricht convergence criteria by financial market participants, particularly those active in advanced industrial nation government bond markets. Following this, I discuss the development of the Maastricht convergence criteria and investigate the political nature of this process. I conclude with an examination of the implications of my study for government-financial market relations.

I. International Financial Market Activity and Key Indicators

The Use of Key Indicators

When making cross-national asset allocation decisions regarding government bonds, financial market participants are concerned with default, inflation, and exchange rate risk. 1 In assessing these risks, market participants employ a set of “key indicators," which are seen to provide a sufficient approximation of risk. The number and range of indicators employed by market participants determines the breadth of the financial market constraint.

If financial market participants rely on a wide range of indicators - if, in effect, they use near-full information rather than information shortcuts - the market-based constraint will be broad. Market actors may look at inflation; the overall government budget deficit; government spending in education, health care, transport, and social security; and labor market policies in various sectors. Because these market actors are interested in and responsive to a wide range of government policies, governments that want to "keep markets happy" are compelled to accede to market preferences in many areas. On the other hand, if financial market participants rely on a narrow set of indicators - for example, if they consider the overall government deficit, the rate of inflation, and government debt, but not the breakdown of government spending or the content of supply side policies - governments that want to "keep markets happy" need only conform to market preferences in this narrow set of areas. In other areas, governments may pursue domestically-determined - and perhaps divergent - policies without fear of negative market response.

With respect to OECD nations, market participants tend to rely on a narrow set of indicators as "information shortcuts." In developed democracies, the potential range of policy outcomes is somewhat narrow: although a social democratic government may be elected in Germany, it is very unlikely that it will increase marginal tax rates to 70% or that it will institute capital controls. It is even less likely that the regime type will shift from democratic to authoritarian. In a sense, the label of "developed nation" is itself an information shortcut: it suggests a degree of certainty regarding government policy, and a narrower range of possible policies. Although investors are concerned about government policy outcomes in developed nations, and are aware of variance in outcomes among this group of countries, they trust the governments to a considerable degree: they find most government statistical releases to be reliable, and they do not worry about the ways in which governments spend their money. Therefore, investors find it acceptable to use a small set of indicators when making investment decisions. For example, when making allocation decisions across a set of developed country government bonds, market actors might look only at inflation levels and government deficit/GDP levels, rather than at how government spending is allocated among consumption, education, and infrastructure. Market actors could rely on a broader set of indicators, and doing so might marginally increase the quality of their asset allocation decisions, but actors conclude that the potential improvements in performance are not justified by the certain increases in information processing costs.

In order to confirm the use of "information shortcuts" and to identify the indicators used by market participants in making asset allocation decisions, I utilize interviews with financial market participants in London and Frankfurt, conducted from January through June 1997. 2 These individuals are active in fixed income (government bond) and equity markets, and make “medium to long-term” decisions and recommendations regarding asset allocation. Most are employed by large investment firms or institutional fund managers, and they distribute investment portfolios across a range of countries, including all OECD and a few emerging market nations. In assessing this evidence, I focus on fixed income, rather than on equities. 3

Interview evidence confirms the reliance of market participants on a few “key indicators.” Because of the nature of the assets, and because of concerns regarding performance relative to other market participants, government bond market participants act as a very homogeneous group. They focus on a similar set of "information shortcuts" when making asset allocation decisions (Interviews 21, 41, 47). This behavior is consistent with a "strong and narrow" depiction of the financial market constraint. When dealing with developed democracies, market actors forcefully demand particular values on key variables, but the number of key variables is small, so that, in other areas, governments remain fairly autonomous from financial market pressures. 4

B. The Identity of Key Indicators

Given the pervasive use of information shortcuts by market participants, a central question regards the identity of these shortcuts: what sort of information do the shortcuts include? Interviews with market participants provide a wealth of information with respect to this issue. When government policy is employed in the decision-making process, it is usually policy outcomes, rather than policy outputs, that matter to market actors. Market actors, however, do believe that a relationship exists between outputs and outcomes, and sometimes react negatively to outputs which are perceived to produce sub-optimal outcomes. For example, in December 1997, the Japanese government announced a tax cut, aimed at stimulating the sagging Japanese economy. Initially, financial markets responded positively to the announcement with advances in the Nikkei equity index and the value of the yen (Financial Times, December 16 and 17, 1997). In the next few days, however, these gains were reversed. Market actors expressed concerns that the tax cut package lacked "enough timely concrete measures to convince markets that there is no risk of a serious economic downturn in Japan," that it was a "muddle-through approach" or a "sticking-plaster solution," and that "none of this will make a difference" to rates of economic growth (Financial Times, December 17, 18, and 22, 1997). The general sentiment was that the policy output (an economic package including $6.5 billion in tax cuts) would not produce the desired policy outcome (stimulation of the Japanese economy).

When examining key indicators, bond market participants' consideration of government economic policy is quite limited. The most widely-used information shortcuts are a set of "big numbers" outcomes: the government deficit/GDP ratio, the rate of inflation, and (sometimes) the foreign exchange rate and the government debt/GDP ratio. Because bond market actors worry about government incentives to default or inflate in response to high levels of debt, the total amount of government borrowing (given by the deficit for any one year, and by the debt for total accumulated borrowing) is the most important fiscal indicator (Interviews 6, 7, 16, 18, 39, 41; Borio and McCauley, 1996). 5 Therefore, this set of aggregate indicators is assumed to capture default and inflation risk and, therefore, to predict relative bond performance. Market participants presently have well-defined preferences regarding these criteria: they want inflation rates of less than two percent and they want low budget deficit/GDP ratios, preferably of less than three percent.

Market actors care about large shifts in government policy affecting outcomes on these key indicators, but not about other policy shifts or the "politics" associated with other areas of government policy. On the other hand, as a result of market actors' short time horizons, very few market participants examine how governments allocate their spending across areas, or even look at the total size of government. To a great extent, if governments are able to finance expenditures via revenue (rather than via borrowing), markets are quite unconcerned about the size of government. "As long as the government is able to pay for what it spends, we do not really care - within broad limits - how much they spend" (Interview 3). Or, "the most important thing is how much governments borrow. The size of government only matters when government is so big as to be a burden" (Interview 41; also Interviews 9, 17). Furthermore, "these things [government spending across issue areas] have implications for long-term growth, but we could sell our assets tomorrow" (Interview 12). "We don't care about where the deficit comes from; the bottom line is that the deficit has to be refinanced by the bond market" (Interview 11). 6

Ultimately, bond market participants “don’t care about the micro-management of the economy” (Interviews 39, 40). A few market participants state that they prefer cuts in government spending to increases in taxation, 7 because spending cuts are a “more certain” revenue source. 8 But, when pressed on this issue, they admit that "at the end of the day, even the 'very long term' people are concerned with the size of the deficit, rather than how the government finances its spending" (Interview 8). Ultimately, if domestic constituents prefer and are willing to fund larger public sectors, governments are not punished by markets for acceding to this demand (Interview 36). "The market controls the big picture, but the government has a lot of discretion in how it spends" (Interview 19). This picture is consistent with recent work by OECD economists, which suggests a strong relationship between deficit levels and government bond rates, but almost no relationship between government spending by area (transfers vs. investment, for example) and government bond rates (Interview, Kennedy, OECD Economics Department; Orr, et al., 1995).

II. How the Rules Matter: the Role of the Maastricht Criteria in International Financial Markets

A. Maastricht and Market Activity

The use of the above "key indicators" is strongly influenced by the development and operation of the Maastricht convergence criteria (see Table 1), particularly the government deficit criterion. Interview evidence suggests that the convergence criteria for EMU serve as a focal point (Schelling, 1960) for market participants. Market actors' use of these indicators, which include "reference values" for sustainable convergence, represents not merely an extension of the "big numbers" criteria, but a change in the way the "big numbers" are evaluated. The criteria make the targets for government policy outcomes more explicit, and therefore, make the "big numbers" constraint sharper - violations are more obvious and significant, and market actors respond to changes in government deficits in light of the Maastricht deficit limit of three percent of GDP. "The [deficit] constraint itself is not entirely new, but the criteria provide a common language for market actors" (Interview 24). "The convergence criteria are a much stronger influence on governments than financial markets alone" (Interview 16; also, Interview 48).

Prior to the mid-1990s, market participants took a "less is better" view with respect to inflation and budget deficits. Four percent is better than five percent, and five percent is better than six percent (Interview 11). There was no clear target under which governments were expected to fall, and no deadline before which they were to do so (Interviews 24, 45). The Maastricht Treaty and its Protocols (the Treaty on European Union, or the TEU), however, offer explicit recommendations (in the form of "reference values") and timetables for potential EMU members: a government budget deficit no greater than 3% of GDP, accumulated government debt of no more than 60% of GDP, and an inflation rate tied to that of the three lowest inflation nations, all by the end of 1997. ). "Ten years ago, the market liked low inflation, low deficits, and good growth. Now we pay attention to specific numbers" (Interview 42). The Maastricht recommendations serve as a specification - and a dichotomization ("above or below three percent") -- of an otherwise fuzzy concept. The TEU demands "sustainable convergence" as a condition for EMU, and market participants look to the "three percent deficit" value as a specification of "convergence" (Interview 22; UK Treasury Official, Europe Division). Since the mid-1990s, these criteria have been "the rules" for market decision-making. In early 1997, for example, government bond market participants watched the Italian government's actions very closely; every move was analyzed according to "will this get them below three percent or not?" (Financial Times, November 19, 1996)

A central reason for market participants' use of the Maastricht criteria as a focal point, of course, is that governments use the criteria. The bond market attempts to predict who will join EMU in 1999. Because performance on the criteria seem to form the basis for governments' decisions regarding EMU, market attention to the criteria is not surprising. In this sense, market actors care not so much about the precise nature of the criteria, but about their use by governments. "There is no economic justification for the criteria, but we look at them because governments look at them" (Interview 11). As the "political" dimension of the EMU decision becomes more salient, the criteria will be less important to market actors (Interviews 8, 17, 21; Interview with Senior Official, European Monetary Institute; HSBC James Capel, 1996, p. 19). A May 1997 headline noted that "the criteria are crucial but ultimately judgments will be political" (Financial Times, May 28, 1997). Additionally, adherence to the Maastricht criteria is interpreted by market actors as a political signal of governments' resolve and credibility: if a government is strongly committed to the single currency, it will find a way to meet the deficit criteria. If a government is unable to meet the three percent criterion, there is reason to doubt its future commitment to EMU. In the words of the European Commission (1995), "Compliance with the criteria is not simply a test of qualification for membership but a lasting commitment by the Member States to adhere to an economic policy conducive to sustainable growth, employment, and the maintenance of purchasing power" (p. 5).

And given that markets attempt to predict which nations will join EMU, we can expect them to pay attention to any factors which might influence the membership decision. These include not only the convergence criteria, but also events which may affect performance on these criteria. In general, current market actors pay little attention to government partisanship (contrary to Kurzer, 1993; McKenzie and Lee, 1991). Rather, they are concerned "about policies, not politics" (Interviews 4, 5, 15, 37 and 45). Because it is what governments do, rather than the labels attached to their parties, that matters to market actors (Interview 11), elections usually do not create medium or long-term changes in financial markets. But there are exceptional cases in which market participants focus on elections; this occurs when government change portends significant policy shifts, such as a movement away or toward EMU (Interview 12). The French elections of May/June 1997 are one illustration of the extent to which the broader issue of EMU is important to market participants in the late 1990s. Market participants worried about the effects of a Socialist victory on France's budget deficit, particularly in light of its attempts to meet the EMU criteria (Interviews 13, 14). Many believed that the Socialists would act differently than the Conservatives: that they would not push through tough spending cuts, that they might enact further unemployment reliefs, and that, ultimately, they might prevent France from meeting the EMU criteria (Financial Times, May 30, 1997; May 31, 1997; June 10, 1997). 9

On the other hand, if the convergence criteria do not appear to be a part of the political process surrounding EMU, we might expect market participants to cease paying attention to them. Many market participants see no reason to set the allowable level of debt at 60% of GDP; several suggest that higher levels are sustainable, while others look at trends in the debt, rather than its actual level. As it became evident to market participants that the "strict" debt criterion (that is, 60%, rather than "sufficiently diminishing and approaching the reference value at a satisfactory pace") would not be a barrier for entry into EMU, market participants also began to ignore the criterion. "The 60 percent criterion has been dropped by the EU, and therefore, largely by markets" (Interview 35; also Interview 16). 10 As with pre-Maastricht deficits, market participants prefer lower debt levels to higher debt levels, but they do not rely on a specific “focal point” amount when judging levels of debt as acceptable or unacceptable. At the same time, though, market actors cling to three percent as a fiscal deficit information shortcut, perhaps following German finance minister Theo Waigel's insistence that "three percent means three percent."

The EU-politics dimension, however, is not the only facet of the Maastricht criteria's use. The criteria have gained "independent" status -- that is, market participants routinely evaluate non-EU states according to the Maastricht criteria: "There has been no conscious attempt to apply the Maastricht criteria to non-EU nations, but it happens. Canada is in really good shape; she would qualify for EMU" (Interview 12), or "the US deficit isn't really much at all, when you use the Maastricht criteria" (Interview, Monetary Affairs Division, Bank for International Settlements).

Finally, the convergence criteria are used widely by market participants despite the fact that many market actors find the criteria objectively flawed or somewhat arbitrary: there is no good, objective reason "to use three percent" (Interviews 8, 12, 24; also European Commission official, DG-II), "to cast 60 percent as an appropriate level of debt" (Interview 41), or "to make no allowance for cyclical variations in the deficit" (Interview 45). Others point out that "it is the convergence of macroeconomic factors, not the level at which they converge, that is important. If all have deficits of four percent, they should move forward" (Interview 6). But, because these criteria have become a focal point for market participants, they remain important. And the widespread use of these criteria as a focal point strengthens the financial market constraint: all market participants employ the criteria (at least the deficit criterion) and respond to them in similar fashions.

B. Implications of the Criteria's Role in Private Markets

The role of the Maastricht criteria in modifying the "big numbers" shortcut, in evaluating non-EU and EU nations alike, and in strengthening the current financial market constraint for developed nations offers an important lesson regarding the criteria: by establishing the criteria with which they later are judged, governments are able to influence their relationship with the international financial market. "Governments are constrained by the Maastricht convergence criteria, but to some extent, it was governments that created the rules for Maastricht" (Interivew 10). And, "governments depend on markets where they have bound themselves to do so" (Interview 37).

The formulation of the Maastricht convergence reference values (as part of the treaty-writing process) was by no means a purely technical exercise. Rather, the criteria represent a "rule" by which the "government-market" game is played. Governments that later adhered to the rules, such as Italy and Spain, quickly gained credibility with financial markets, much more quickly than they would have done independently of the move toward EMU (Interviews 5, 11, 18; Financial Times, May 20, 1997; also see DeGrauwe, 1992, p. 54). 11 For example, in the fifteen months beginning in January 1996, the interest rate spreads between German and Italian ten-year government bonds fell from approximately five percent to just over two percent (Financial Times, March 17, 1997). This effect also operates in the reverse direction: any speculation in the financial markets regarding a delay of EMU generates a negative reaction in the Italian bond market, as investors worry that, without EMU, "the Italian government will loosen its policy of fiscal restraint" (Financial Times, August 14, 1997). Therefore, despite the constraints that the Maastricht rules represent, governments are able to influence the character of these rules and to profit from adherence to them (Interview, European Commission DG-II). Hence, we see a subtle means of government influence on the nature of externally-imposed constraints.

More broadly, when selecting key indicators, market participants sometimes look to governments. "One reason that indicators become important is that governments pay attention to them. If the government makes a point of certain statistical releases, then markets are going to pay attention to those figures. That was a big motivation for looking at balance of payments data for many years" (Interview, UK Treasury Official, Europe Division). Similarly, "part of the ERM crisis was that countries set themselves up. They gave the markets indicators to worry about, and markets attacked when these indicators went wrong" (Interview, European Commission official, DGII). Moreover, when governments cease to pay attention to certain indicators, as they did with the Maastricht debt guideline, markets also lose interest in these indicators (Interview 16).

III. Making the Rules: the Development of the Maastricht Criteria

Given the importance of the Maastricht convergence criteria, especially the fiscal criteria, to financial market activity and, therefore, to governments, it is useful to examine the development of these criteria. Why were these particular indicators chosen as the basis for EMU membership? Given that many market actors see no economic rationale for these particular numbers, why were these reference values for debt and deficit chosen?

Economic and monetary union has been discussed in Europe since the late 1960s. In 1969, the Barre Report, issued by the European Commission, argued for a full EMU. The Werner Report, issued in October 1970, discussed a possible movement toward a single currency; in response the Council of Ministers adopted a plan to achieve EMU by 1980 (George, 1996). These attempts, however, were put aside after the breakdown of the Bretton Woods exchange rate system and the first oil shock. Of course, the launching of the European Monetary System and the ecu currency in March 1979 also reflected a desire to move toward a single currency.

Beginning in 1987, the French government called for greater European cooperation in monetary affairs. 12 In its 1988 meeting at Hanover, the European Council addressed the objective of EMU, and commissioned the Delors Committee, headed by Jacques Delors and comprised of the governors of European central banks, to study the issue of EMU and present a plan to move toward EMU. This report, issued in April 1989, contains most of the blueprint for EMU and was endorsed in December 1989 by the European Council. With this, the movement toward the contemporary incarnation of EMU began in earnest. 13 In October 1990, the European Council adopted proposals for the specifics of EMU; these informed the 1991 Intergovernmental Conference (IGC) as well as the Maastricht Treaty. With the formal signing of the Maastricht Treaty in 1992 and its ratification in 1993, EU members entered the first stage of EMU.

That the broad outlines of these efforts at EMU reflect political pressures and motivations is widely acknowledged (for example, see Feldstein, 1997; Garrett, 1993; Sandholtz, 1993). In terms of pure economic rationale, it is difficult to argue that the EU constitutes an optimum currency area (DeGrauwe, 1992; Gros and Thygesen, 1992; Sala-I-Martin and Sachs in Canzoneri et al., 1992), or that Germany would accede to EMU for purely economic reasons (Grieco, 1993). A movement toward a single currency entails some surrender of autonomy, particularly over exchange and interest rate policy. In return, EMU removes intra-EU currency fluctuations, eliminates the associated transaction costs, and facilitates the full realization of a Single Market (Delors Report, para. 26; Eichengreen, 1992; Padoa-Schioppa, 1994).

But, even if member states agree on the broad objective of EMU, many issues remain: will a new currency replace national currencies, or will some sort of parallel currency exist? 14 Will future members be required to meet certain conditions prior to joining the currency area, or will convergence of economic performance come as a result of the movement to a single currency? If conditions for entry into EMU are required, what will the conditions be? At each of these levels, we see political maneuvering, as gains and losses for certain nations and groups are identified. In describing progress toward EMU, economist Tomaso Pado-Schioppa points out that the role of the economist and the expert is to look for the technical solutions by means of which to overcome this fundamental institutional and political problem. But if it is not clearly recognized that these technical issues are just secondary aspects of a basically political problem, the essential point will be missed (1994, p. 85).

But, as the level of the decision drops, from the general movement toward EMU to the identity of convergence criteria, the attention given to political factors in academic literature and popular media also declines. In this section, I first discuss the political debates surrounding the "larger issues," then turn to main topic of this paper - the reference values of the fiscal convergence criteria.

A. The Transition: "Monetarists" vs. "Economists"

The Delors Report outlines a three-stage progression to EMU: in Stage One, EC members outside the EMS join the system, and a single, narrow fluctuation band is employed. In Stage Two, members' fiscal policies are closely coordinated. Finally, in Stage Three, members lock their currencies together. While careful to set out each of these stages, the Delors Report devoted much more attention to the end stage - the realization of EMU - than to the path (the "second stage"). In the design of Stage 2, the Delors Report left room for political maneuvering and bargaining. "After the publication of the report, the design of the transition soon emerged as the most difficult and unresolved problem of the Treaty" (Bini-Smaghi et al, 1994, p. 9). This was an issue left more to "political discretion and bargaining" than to technical analysis (Fratianni et al., 1992, p. 6).

At the root of this ill-defined path was a tension dating back to the Werner Report (George, 1996): was macroeconomic convergence a necessary precondition for EMU, or a consequence of EMU? 15 The former view, labeled as the “economist,” “behavioralist,” or “gradualist” approach, maintains that convergence of monetary policy, fiscal policy, standards of living, and wages is a prerequisite for EMU, and that "convergence had to be measured according to precise quantitative criteria" (Bini-Smaghi et al, 1994, p. 12). Similarly, the Werner Report described the first stage of EMU as one that "would focus on the coordination and convergence of monetary and fiscal policies" (Kenen, 1995, p. 5; Gros and Thygesen, 1992). The latter perspective, the "monetarist," "institutionalist," or "shock therapy" approach, holds that institutional and market-based changes - including the adoption of a single currency and the creation of a supranational monetary authority - will lead to convergent behaviors and policy outcomes (DeGrauwe, 1992; Gros and Thygesen, 1992, p. 272; also Interview 41). The Delors Report navigated a middle course between these two positions. It observed that "parallel advancement in economic and monetary integration would be indispensable in order to avoid imbalancesÉ but perfect parallelism at each point in time would be impossible and could even be counterproductive" (Commission, para. 42). In this spirit, the Delors Report did not mention specific convergence conditions.

The use of convergence criteria in Stage Two was a revolutionary idea, in terms of past efforts at monetary union: no other such efforts required prior macroeconomic convergence (Crowley, 1996; DeGrauwe, 1996; IMF Survey, Sept. 9, 1996, p. 280). In general, Germany and other northern European nations advocated a long Stage Two, with sufficient convergence prior to EMU; France and the southern European countries favored a shorter Stage Two, and an EMU preceding full convergence (Milner, 1997). Additionally, some - including the British government -- suggested that market discipline would provide sufficient pressures on government policies: governments with "excessive" inflation, deficits, or debt would be judged by markets as unfit for EMU, would be punished with higher interest rate premia and, as a result, would modify their policies in a convergent direction (Bayoumi et al., 1995, pp. 1047-1048; Padoa-Schioppa, 1994). But the Delors Committee found market discipline to be too loose. 16 In the words of a Commission official, “market discipline is not consistent; they still threaten governments, but punishment often is not instantaneous. Markets require a college-level ethics, not kindergartner behavior, where you’re punished immediately” (Interview, European Commission Official, DG-II).

In the period immediately following the Delors Report, the EC moved increasingly in the “economist" direction (Padoa-Schioppa, 1994, pp. 198-199). The view expressed by Germany and France was that solid budgetary discipline was essential for the success of EMU and should be required for entry into Stage Three (Bini-Smaghi et al., 1994; Buti et al, 1997, p. 2). Likewise, in August 1990, the European Commission stated that entry into the third stage of EMU should depend on macroeconomic convergence. Additionally, in the Monetary Committee during 1990 and 1991, the notion that objective and quantitative convergence requirements had to exist slowly gained ground, with strong backing from the Bundesbank-inspired German and Dutch delegations (Crowley, 1996, p. 42). The Committee's July 1990 report to the Ecofin Council called for objective criteria, applied "not in an automatic way, but with the exercise of judgment" and including price stability, budgetary discipline, participation in EMS, a positive market assessment of the sustainability of convergence, and completion of the internal market" (Bini-Smaghi, et. al, 1994, pp. 21-23; also see Delors Report, para. 30). These recommendations closely foreshadow those found in Articles 104 and 109 of the TEU.

Before examining the development of these criteria, it is useful to consider the political motivations underlying the choice of an "economist" over a "monetarist" strategy. For what political reasons is it important to have ex ante convergence criteria? One is an inherent suspicion regarding the efficiency of financial markets: although markets do discipline governments, they do not always do so in a continuous fashion (for example, see Missale and Blanchard, 1994; Economist, October 7, 1995; International Monetary Fund, 1997, p. 56). Governments have a "good deal of rope with which to hang themselves" (Interview, Senior Economics Adviser, UK Foreign and Commonwealth Office), and market activity fails to discipline governments sufficiently, at least in the short and medium runs. "Éexperience suggests that market perceptions do not necessarily provide strong and compelling signals and that access to a large capital market may for some time even facilitate the financing of economic imbalancesÉ.The constraints imposed by market forces might either be too slow and too weak or too sudden and disruptive" (Delors Report, para. 30; also Kenen, 1995, p. 93).

Another motivation for ex ante convergence concerns a desire to avert political tensions after EMU. The Delors Report notes that, within the EMS, a lack of "sufficient convergence" of fiscal policies placed undue pressures on monetary policy and, therefore, was a source of political tension among members (para. 5). Therefore, "greater convergence of economic convergence is needed" (para. 11). This convergence could also serve to reassure markets that the EMU project - and its members' economic policies - was credible (Holzmann et al, 1996, p. 34). "All the governmentsÉneeded to transform their reputations from macroeconomic laxity by attaching themselves quickly to an international institution with strong counterinflation credentials" (Garrett, 1993, p. 114).

A final motivation is a concern regarding the commitment of other member states to the EMU project (or at least a determination to require states to demonstrate their commitment). If no costs were associated with joining EMU, all states would join; but these states might not be willing to bear future costs and could destabilize the entire project. Hence, the existence of some set of conditions raises the bar for admission and compels states to demonstrate a commitment to Community norms. In this way, EMU membership "might be intended as a reward for good behavior" or a "test of will" for members (Frankel, commenting on Buiter et al, 1993, pp. 94-96).

Defining Convergence

With broad agreement on the desirability of the convergence criteria, the character of these criteria became the next important issue (Sandholtz, 1993). The Delors Report offers no specific criteria, but suggests the use of "binding rules and procedures for budgetary policy," including upper limits on budget deficits and "the definition of the overall stance of fiscal policy during the medium term" (para. 33). In the ICG, "convergence" was defined as not only similar performance across nations, but also as "sustainable" and "satisfactory" performance (Crockett in Steinherr, 1994, p. 171); at the political level, the emphasis was on fiscal discipline rather than on fiscal coordination (Padoa-Schioppa, 1994, p. 168).

The convergence criteria were central to the IGC from April 1991. Only Greece objected to the inclusion of any sort of rules concerning government deficits (Crawford, 1993, p. 138). Otherwise, there was a "consensus that EMU would only be sustainable if underpinned by a considerable degree of economic convergence" (George, 1996, p. 225). After five months of negotiations, the Monetary Committee proposed a set of three convergence criteria, which were not to be used "mechanically:" inflation (perhaps no more than 1.5% greater than the two or three best performers); participation in the ERM, and no excessive deficits (defined in terms of both deficit and debt). The long-term interest rate criterion was added soon after, at the insistence of Germany (and despite the objections of France and Italy) (Bini-Smaghi et al, 1994). 17

This phase of the criteria-development process also included a debate regarding where the criteria should appear: in the Treaty (and, therefore, only modified with a unanimous vote), or in secondary Community legislation (and modified by majority vote). Belgium, Italy, Portugal and Greece wanted the criteria to be included in secondary legislation; they worried that the criteria would be too strict and difficult to change. Germany and the Netherlands, with the support of France, advocated Treaty-based criteria, which would prevent political influence on the convergence standards. The outcome reflected this view: the criteria could be modified only by a unanimous vote of the Council (Bini-Smaghi, 1994).

It is quite possible to justify the inclusion of the inflation criteria in the Treaty on economic, as well as political, grounds: members hoped that financial market participants would expect a European-wide inflation level close to the German level, rather than a level that was an average of all participants (e.g. Germany and Italy). Influencing market expectations of future inflation entailed changing current inflation levels (DeGrauwe, 1996; Fratianni et al., 1992, p. 21) and expectations regarding future attempts to monetize outstanding debt. The inclusion of a long-term interest rate criterion in the set of convergence conditions also reflects an attention to financial market judgment (although the interest rate criterion receives much less attention than the fiscal criteria).

The inclusion of fiscal convergence criteria in the Maastricht Treaty, though, seems driven largely by politics, rather than by economics (Wyplosz, commenting on Buiter, et al, 1993, p. 97). While some economists argue that fiscal restraints provide an effective means of reducing budget deficits (Eichengreen, 1992), others maintain that the existence of any set of fiscal criteria create incentives for sub-optimal behavior: governments will attempt to satisfy the criteria in ways that are politically expedient (e.g. cutting public investment spending) rather than in ways that produce long-term structural reform (e.g. undertaking pension reform) (Holzmann et al, 1994; Masson in IMF Survey, September 9, 1996, p. 280).

The particular criteria are not those that traditional economics would recommend: in thinking about optimum currency areas, economists would emphasize micro-side factors, such as wage flexibility and labor mobilityÉ.Economists might also stress the need to fiscal flexibilty, in order to deal with shocks (especially when monetary policy is limited) (DeGrauwe, 1996).

On the other hand, the Delors Report calls for "precise quantitative guidelines for fiscal policy" and "concerted budgetary action by member countries" (para 51). No explicit rationale for fiscal criteria is given, however, in the Delors Report or in subsequent European Commission publications (Holzmann et al, 1996, p. 26).

In the political discussions of the early 1990s, member governments repeatedly expressed concerns about "fiscal irresponsibility." High levels of government debt and profligate fiscal policies could lead to several situations: states could raise interest rates for the entire currency area (Holzmann et al, 1996), crowd out private investment, create political pressures on the European Central Bank for community-wide inflation (Crockett in Steinherr, 1994), or force other members to enact a debt "bail-out" (Buiter et al, 1993; Crowley, 1996; Eichengreen, 1992; DeGrauwe, 1996). Although many commentators have argued that the concerns are unfounded - for example, market actors are able to differentiate between governments, and would only charge high interest rates to those with unsustainable fiscal policies (Buiter et al, 1993, pp. 78-79) - they were salient for EU governments. The "no bail out" clause appears in Article 104b of the Maastricht Treaty, and limits on deficit and debt levels provided a means to render this clause credible (Kenen, 1995, p. 94).

Finally, governments may have supported the implementation of a set of criteria for domestic political reasons. The criteria allowed states to undertake unpopular economic measures without accepting full responsibility for these (George, 1996, p. 222) 18 , and also to gain credibility with international financial markets. 19 Several government officials point to the “scapegoating” role of the Maastricht criteria (also see Andrews and Willett, 1997; Buiter et al, 1993, p. 85; Pierson, 1996). “Governments already had to deal with high deficit and debt levels, but the EMU criteria forced them to do so dramatically” (Interview with senior Bundesbank official). “There was a common consensus that a government debt/GDP ratio of more than 50% was to high. The current round of cuts had to be made, with or without the Maastricht criteria. The criteria should be seen as an instrument by which states can get their policies in order" (Interview, senior German Economics Ministry official; also International Monetary Fund, 1997, p. 66). This returns to a theme found at all levels of the EMU project: although governments agreed to constraint themselves as part of the EMU effort, they also gained political "space" in some areas.

What Kind of Convergence?

The next decision facing member states was how to define the fiscal criteria - to look only at overall deficits, or to look at particular types of government spending; to assign a similar rule across countries, or to allow for differences in levels of economic development and position in the macroeconomic cycle. 20 The Delors Report was silent on this issue (Kenen, 1995). Members decided to look at overall deficits and nominal gross government debt, 21 with the same reference value for all countries and all years. This decision was somewhat contrary to the suggestions of economists, as well as to some governments' proposals. For example, some economists expressed concern that the criteria were defined too crudely: they do not differentiate between the structural balance (which measures what governments are doing with taxation and spending) and the cyclical balance (which measures the effects of the business cycle on government outlays) (Brian Redding, Financial Times, April 5, 1997). Therefore, a government facing a recession is subject to the same rules as a government in an expansion; the criteria may impede the functioning of automatic stabilizers (Kenen, 1995, p. 107). 22

In addition, the criteria do not separate the primary deficit (which excludes interest payments on government debt) from the overall deficit, leaving high-debt governments more constrained in the non-interest payment areas of fiscal policy (Gros and Thygesen, 1992; Weber in Torres and Giavazzi, 1993, p. 90). Furthermore, given that, in the early 1990s, governments were in very different fiscal positions - Italy and Belgium were facing near-unsustainable levels of debt, while France and the Netherlands were not - why should all be subject to the same rules? (Buiter et al, 1993, p. 69; Gros and Thygesen, 1992; Kenen, 1995, p. 93). Perhaps the criticisms of the fiscal criteria reflect the fact that "it is not at all simple to devise rules that are sufficiently transparent and intuitive to be incorporated into a draft minute for agreement by Finance Ministers" (Currie in Canzoneri et al., 1992, p. 143). 23

During the IGC, some governments were in favor of "differentiation of deficits." The Germans and the Dutch suggested a "Golden Rule," in which budget deficits would be permitted only for capital (and not for current) outlays. The German government already was constrained domestically by this rule, in Article 150 of its constitution. But the use of an explicit Golden Rule was rejected, on the grounds that it leads to counter-cyclical government spending (Bini-Smaghi, 1994; Buiter et al., 1993). At the same time, France proposed the use of cyclically-adjusted deficits, but this idea was rejected on the grounds that cyclically-adjusted deficits were difficult to measure accurately. The final decision on this issue, made by the members in April 1991, was to institute a static limit on deficits and a dynamic limit on debt. The Draft Treaty directed the Commission to consider the deficit/GDP ratio, "which shall not exceed a ceiling" and the ratio of debt/GDP, which "may only exceed a reference value if the ratio is sufficiently diminishing and steadily approaching the reference value" (Treaty on European Union, Articles 104 and 109). 24 It seems that a key motivation behind this outcome – behind the choice of very aggregate indicators and a lack of differentiation – was a concern about the difficulty (and potential political conflicts) associated with measuring fiscal aggregates (Crockett in Steinherr, 1994, p. 179; Gros and Thygesen, 1992, p. 284). 25 In retrospect, this concern probably was well placed: the tension over "one off measures" and "creative accounting" that became apparent in 1997 may have been worse with a more differentiated deficit rule.

This leads to a final question, the importance of which is revealed in the "focal point" use of the Maastricht deficit criterion by international financial markets: why three percent and sixty percent? The decision to include fiscal convergence criteria as a condition for participation in EMU left unresolved the particular reference values for these criteria. In their arguments to include fiscal criteria, member governments had suggested that all nations should have similar overall fiscal policies and "sustainable" levels of debt, but they had not assigned particular values. A motivating idea behind the selection of the criteria was that Germany and the other nations needed to converge, in terms of macroeconomic performance (Garrett, 1993). But "convergence" could have been on any level, within some reasonable range. Fiscal convergence could have entailed deficits of one percent, four percent or six percent of GDP, and debt of forty, seventy, or ninety percent of GDP. Why, then, when choosing the reference values in December 1991 (Garrett, 1993, p. 106), did the members settle on three percent and sixty percent of GDP? Moreover, given that the criteria became a focal point for market participants, whose interests were best served by this particular choice of criteria?

Again, politics, rather than economics, may have motivated the choice of reference values for the fiscal criteria. Given that traditional economic texts did not advocate the use of prior convergence criteria, they also were of little help with respect to reference values. The one piece of guidance on reference values provided by economists is that debt and deficit values should be consistent with one another; assuming a particular rate of nominal economic growth, the deficit reference value should allow for a stabilization of debt at the debt reference level. So, a debt/GDP ratio of 60 percent is stabilized with a government deficit/GDP ratio of three percent and a nominal growth rate of five percent. 26 But, as long as the deficit and debt numbers are mutually consistent, any reference values could be chosen: “there is no theory at all behind the numbers; any numbers could have been used” (Interview, European Commission Official, DG-II; also Buiter et al, 1993, p. 63). The real question, then, remains why these numbers were chosen. 27

The rationale behind the convergence criteria does not stem from financial market pressures: although financial market participants approve of the timetables and definite goals embodied in the Maastricht Treaty, they did not demand particular values on the fiscal criteria (Crowley, 1996). Rather, four possible explanations of the reference values have been offered: they reflect the average level of government debt in the EC in 1991, when the criteria were selected; they reflect the average of government capital-expenditure deficits; they reflect a deliberate effort to create "high hurdles" and, therefore, to either exclude or to force strong commitments from some potential members; or they reflect a deliberate effort by governments to tie their hands. The latter two explanations suggest that the political importance of the reference values probably was realized in 1991. On the other hand, the "average" explanations suggest that policymakers may not have considered the importance of their choice: numbers that were arbitrary when chosen become quite important when used for political and financial-market decisions. Of course, these explanations are not mutually exclusive, and the choice of reference values may reflect some combination of political maneuvering and technical arbitrariness.

The selected reference values reflect the values proposed by the Netherlands delegation, with some added allowances for flexibility (Ross, 1995, p. 187). During negotiations regarding the reference values, the debt criterion was selected first. Observers note that, after a good deal of discussion, "the choice finally fell on sixty percent, which was more or less the community averageÉ.Contrary to what is often stated, sixty percent was not considered to be the limit of acceptability for the debt but, rather, a threshold beyond which the constraint on its rate of change becomes relevant" (Bini-Smaghi, 1994, p. 29). The average debt level in 1991 was 61.7%, while many economists point out that "the reference value for debt is far lower than the numbers usually cited in the analytical literature on the solvency problem" (Kenen, 1995, p. 100, fn. 36). We could assume, then, that value chosen was somewhat arbitrary (Interview with European Commission Official, DG-II) or that policymakers chose sixty percent because it was an existing "political" focal point. And perhaps some member governments realized that by including a "sufficiently declining" proposition, they would not be required to meet the 60 percent reference value.

If we assume an annual nominal growth rate of five percent, the choice of a sixty percent debt reference value implies a deficit value of three percent. A deficit/GDP limit of three percent then serves as an "early warning indicator" for debt: as long as deficit/GDP remains at or below three percent, debt/GDP ratios will converge toward sixty percent (Gaspar in Torres and Giavazzi, 1993, p. 40). But policymakers did not accept this automatically (Bini-Smaghi, 1994; Kenen, 1995): the Netherlands proposed the three percent value, on the basis of work by the EC's Monetary Committee. Some nations found this limit too restrictive, even though it was consistent with "sixty and five;" they were concerned that, in some cases, the deficit could rise above three percent without undermining monetary stability. For this reason, they opposed a "mechanical" application of the limits (Cosetti and Roubini in Torres and Giavazzi, 1993, p. 47). Italy and France proposed a compromise, with support from Britain: temporary and limited deviations from 3% were allowed, as long as there was evidence of adjustment measures designed to bring deficits back within limit. This, of course, is at the heart of the Treaty's "excessive deficit" procedure. Of course, this description of events leaves open the question of the reference value's source. Although the "golden rule" was rejected as an overall guideline for deficit levels, other observers point out that EC public investment averaged almost exactly three percent of EC GDP during the 1974-1991 period. (Buiter et al p. 63) 28

The "hurdle" and the "scapegoating" explanations suggest explicit political influence on the choice of reference values: in the hurdle case, there was a deliberate effort made - particularly by Germany - to force states to demonstrate their desire to participate in EMU or, at the extreme, to prevent the entry of particular states. Evidence for this argument is mostly circumstantial: it would not be surprising that the German Bundesbank advocated such rules, but the Bundesbank was not - at least directly - party to the treaty (although they did exercise significant influence over the German delegation; see Milner, 1997). Still, some have suggested that the criteria "reflect not economic logic, but a mixture of German horror at the Italian national debt, and Dutch PuritanismÉ" (Buiter et al., 1993, p. 88).

Additionally, fiscal reference values represent a high hurdle for southern European nations. In 1991, only Germany, France, the UK and Luxembourg satisfied the criteria, strictly defined; in 1992, only France and Luxembourg satisfied them. The Italian fiscal deficit was 10.2% of GDP in 1991; that same year, Greece's deficit was 16.2% (IMF, 1997; also Fratianni et al., 1992, p. 9; Gros and Thygesen, 1992). Moreover, although the debt ratio of a country with annual nominal GDP growth of five percent and a three percent budget deficit will ultimately converge to 60 percent, this convergence will require more than a decade for a government with a debt/GDP ratio of 100 percent (Kenen, 1995). It is difficult, then, to imagine some governments meeting these criteria by 1995 or 1997 (Crawford, 1993, p. 253; Letiche, 1993; Sandholtz, 1993; Thygesen in Torres and Giavazzi, 1993, p. 16). This evidence suggests, circumstantially again, that the criteria were an effort to limit EMU to northwest Europe. Detractors of this view would point out, however, that the reference values did not look entirely unrealistic: in 1990, six countries had budget deficits less than three percent, and of the seven countries with debt ratios greater than 60 percent, three were less than 80 percent (Kenen, 1995, p. 101).

The final alternative is that the strict deficit criterion was desirable to governments, not only to Germany and the Netherlands, but to the "Club Med" governments as well. This explains the willingness of southern European nations to accept the fiscal reference values. Perhaps these governments realized that, for domestic reasons, they needed to reduce public expenditure and pubic debt, but did not want to take responsibility for fiscal cuts. By choosing a low deficit level, they bound themselves externally, allowing for domestic scapegoating and enhanced international credibility (see above). This argument is consistent with the fact that many EC governments - not just Italy and Greece - had deficits greater than three percent in 1991; the community-wide average was 4.3 percent. Moreover, interviews with government officials in France and Germany support this interpretation (Interview, French Treasury Official; Interview, Senior Economics Adviser, German Federal Chancellory).

In the end, it is difficult to imagine that governments were not conscious of the significance of the fiscal reference values, both to future governments and financial market participants. Of course, the Treaty allows for some derogation from and political interpretation of the criteria, especially the debt criterion, which permits "satisfactorily declining" debt levels above 60 percent in the face of "good faith" efforts at fiscal consolidation (Artus, 1993; Garrett, 1993; Thygesen, 1996). Again, this is generally consistent with the Delors Report framework, which states that "the situation of each country might have to be taken into consideration" (para. 30). Observers suggest that the vagueness in the debt criteria was intentional (DeGrauwe, 1996, p. 20), perhaps a reflection that sixty percent was an impossible hurdle for some members. There is also an excessive deficit procedure, which describes the circumstances under which deficits may exceed three percent of GDP. The allowance for deficits greater than three percent was not made in the Netherlands Draft, submitted six weeks before the Maastricht conference (Kenen, 1995, p. 100), and is a product of some governments' uneasiness regarding strict fiscal rules: "at the political level, the Treaty might not have been passed if specific rules on stronger fiscal surveillance and sanctions, enhanced streamlining of expenditure and revenue programs, fiscal coordination, or even fiscal federalism had been included" (Holzmann et al, 1996, p. 53).

In practice, then, the debt criterion has been applied loosely, but the deficit criterion has been applied quite stringently (or "blindly and mechanically"), by national governments and particularly by financial market participants. The Commission's 1995 Green Paper calls for strict application of all criteria:

The four criteria are both key indicators of sound economic management and the objective basis for political decision. They must be strictly and fully respected, both before and after the changeover to the single currencyÉ.This is the only was to build Monetary Union on solid foundations and to guarantee its permanence. It is also the only way to provide Europe with that stable macroeconomic environment it needs for faster growth and more job creation (p. I)

Market actors treat the deficit reference value as a dichotomous focal point: either a government's deficit meets the criterion, or it exceeds it. The result, of course, is that, in the run-up to EMU governments have been rather tightly bound - for political and financial-market reasons - to three percent.

This is not to suggest that no reference values should have been selected by EC governments. In terms of convincing financial market participants of the credibility of the EMU project - and in terms of allowing governments "to set the rules by which they are judged" - specific reference values are very useful. "The figures are arbitrary, but they have done a good job of grabbing member states' attention" (Interview, European Commission Official, DG-II). Or, precision may have been overdone in the reference values adopted...[but] any specific numerical value for the ratios of the deficits and debt to GDP may well be difficult to justify analytically, but the consequence of offering no quantitative precision whatsoever would have been to remove the strong incentive to aim for major improvement in budgetary performance now contained in the Treaty." (Gros and Thygesen, 1992, p. 389).

In other words, when financial markets and governments have a clear understanding of the "hurdles" for governments, they can much more easily evaluate government policy in light of these goals. There is a tradeoff, of course: simple rules do not allow for consideration of the complex interactions which characterize fiscal policymaking and fiscal policy outcomes. Moreover, the existence of simple rules creates incentives for "creative accounting" or for "easy methods" of meeting the targets (Holzmann et al, 1996, p. 40; Masson, IMF Survey, Sept. 9, 1996, p. 280). But, despite these problems, a set of specific fiscal rules does "provide reasonable reference points for assessing whether countries are straying too far from their medium-term goal... and thus a barometer of their commitment to that goal" (IMF, 1997, p. 67). 29

Rather, given the role of the criteria in political decisionmaking and cross-national asset allocation, and governments’ ex ante awareness of the role of private markets in evaluating EMU, governments - and political scientists -- should pay closer attention to the selection of specific reference values, rather than treat this selection as a "technical" issue or an "arbitrary choice." It is hardly revolutionary, or even surprising, to argue that the motivations behind the larger EMU project were largely political (Garrett, 1993), but it is less obvious and much less widely acknowledged that the motivations behind the convergence criteria and their specific reference values were political, at least in their ex post effects.

IV. Implications: EU Governments and the Financial Market Constraint in the Future

Most scholars of international and comparative political economy note that governments are somewhat constrained by financial market activities, both in general and in with respect to EMU. The framers of the Maastricht Treaty realized this, and the EMU convergence criteria draw on market activity to suggest who should join EMU. But it is a mistake to overlook the voice of governments in this process, in terms of providing international financial market participants with decision-making rules ("focal points") and in terms of gaining from adherence to these rules (as in the case of Italy's "virtuous circle" of debt reduction).

Because the Maastricht fiscal criteria are approaching (or have reached) the end of their usefulness as focal points to market participants, it is important to consider the future of the financial market constraint for EU nations. As the political side of the EMU decision looms larger, the strict reference values lose meaning, at least as predictors of EMU membership (Interview 29). Therefore, in the post-EMU environment, what government policy measures will serve as "focal points" or "key indicators" for financial market participants?

Within the EMU zone, there will be no cross-national differences in exchange rate or inflation risk for government bonds (although these risks will exist vis-a-vis third countries' government bonds). The key cross-national different will be in default risk (Interview 35), so we can expect market participants to devote the bulk of their attention to this risk. In evaluating this risk, we also can expect market participants to rely on the rules created by governments.

A potential focal point for market participants is the Stability and Growth Pact, agreed upon by the European Council in June 1997. According to the Excessive Deficit provision of the Maastricht Treaty (Article 104c), the three percent threshold can be exceeded without causing an "excessive deficit" judgment if: "(1) the origin of the excess is outside the normal range of situations; (2) the deficit is allowed to remain above 3% for only a limited period of time; (3) the deficit remains close to the reference value." The Stability and Growth Pact adds to Article 104c by (1) setting time limits to the various steps of the Excessive Deficit procedure so as to speed it up and perhaps impose sanctions sooner; (2) defining the meaning of temporariness and exceptionality; (3) and specifying the conditions under which sanctions will be applied and their scale (Buti et al, 1997, p. 3). The Pact promises a means of "keeping governments in line" after 1999 (Interview 35). Financial markets, then, might charge higher risk premia to governments with deficits greater than three percent, unless these deficits come during periods of severe recession. Of course, such a prediction assumes an increasing level of sophistication by market participants, in which they look not only at an aggregate deficit outcome, but at the sources of that outcome.

In the end, there are subtle methods for governments to gain "room to move," despite their decisions to constrain themselves and despite the external constraints stemming from financial market pressures. "Setting the rules," as EU governments did within the Maastricht framework, is one of these methods. Other methods include modifications in the term structure of government debt, where nations with longer term structures tend to have greater autonomy in their relationship with international financial markets and the adoption of credibility-enhancing monetary and fiscal policymaking institutions.

Appendix: Financial Market Interviews

Interview
Number
Firm Date
1 Salomon Bros. 3/5/97
2 Matushita UK 2/25/97
3 Cursitor-Eaton 1/28/97
4 Henderson Inv. Mgmt. 3/21/97
5 AXA Law and Equity 1/27/97
6 Yamaichi Investment 3/26/97
7 Deutsche Bank 4/4/97
8 UBS 1/28/97
9 British Aerospace Pension Mgmt. 1/31/97
10 IBJ International 3/27/97
11 Commerzbank Investment 5/27/97
12 Deutsche Bank Asset Mgmt. 5/22/97
13 J.P. Morgan Investment 5/27/97
14 J.P. Morgan Investment 5/27/97
15 Prudential Inv. Mgmt. 2/11/97
16 BZW (Barclay's) 5/12/97
17 Salomon Bros. 5/9/97
18 Paribas Inv. Mgmt. 3/10/97
19 Nikko Europe 2/13/97
20 Kawasaki (UK) 2/25/97
21 Dresdner Bank 4/8/97
22 Merrill Lynch 4/25/97
23 Institutional Fund Mgrs'Assoc. 2/3/97
24 Deutsche-Morgan Grenfell 2/20/97
25 Honda UK 3/24/97
26 Mikita UK 2/21/97
27 HSBC-James Capel 3/12/97
28 Genesis Inv. Mgmt. 1/20/97
29 Commerzbank Asset Mgmt. 4/16/97
30 Zurich Investment Mgmt. 4/14/97
32 Lombard St. Research 2/19/97
33 Cursitor Eaton 1/28/97
34 Hambros Bank 3/11/97
35 BHF Bank 4/17/97
36 Allianz 4/21/97
37 Deutsche Bank Rsch. 4/9/97
38 Smithers Research 2/19/97
39 Dresdner Bank Rsch. 4/18/97
40 Generale de Banque 5/7/97
41 Deutsche-Morgan Grenfell 2/20/97
42 J.P. Morgan Investment 5/27/97
43 AIWA Industries 3/3/97
44 IBM (UK) 3/5/97
45 AMP Inv. Mgmt. 3/20/97
46 Robert Fleming Inv. Mgmt. 2/21/97
47 Robert Fleming Inv. Mgmt. 2/27/97
48 Deutsche Bank Research 5/21/97

References

Note 1: For further information about the criteria used by financial market participants, or about the “strong but narrow” view of financial market constraints, see Mosley (1998). Back.

Note 2: Throughout the paper, interview subjects are identified by number. A list of subjects’ firms and dates of interview is provided in the Appendix. Interviews with government actors are identified according to agency and nation. Back.

Note 3: Most of those interviewed deal with both equities and government bonds, although national government policies tend to be less important to equity allocation than to government bond allocation (Interview 5, 8, 15, 19, 23, 24, 27, 32, 33, 38). Of course, some heterogeneity exists within financial markets, particularly across types of instruments (Frankel, 1986; Pollins, 1993). Back.

Note 4: On the other hand, those market participants active in emerging markets employ a wider set of information shortcuts: they prefer to consider a wider range of government policies when making asset allocation decisions and express concerns about the “trustworthiness” of emerging market governments. Back.

Note 5: Thorbecke (1993) explores the connection between budget deficits and interest rate levels. Back.

Note 6: A small minority of market actors offers a dissenting view regarding the composition of government spending (Interviews 21, 29). Back.

Note 7: A common phrase among bond market participants is, “as long as the capital gains tax is low…” [we don’t care about other tax rates, to some limit]. (Interview 28) Back.

Note 8: Also see IMF Survey, January 27, 1997. “Budgetary contractions that concentrated on expenditures, especially on transfers and government wages, were more likely to succeed in reducing the public debt ratio than tax increases” (p. 24). Back.

Note 9: These concerns were borne out, if only slightly, immediately after the election: at the beginning of the European Union summit in Amsterdam, newly elected French prime minister Lionel Jospin called for changes in the Stability Pact and a political counterweight to the European Central Bank. Although France's opposition was quelled rather quickly, it suggested to market participants that their worries about French elections were somewhat justified (Financial Times, June 10, 1997; June 17, 1997). Back.

Note 10: Financial Times commentator Martin Wolf explains the abandonment of the debt criteria (Apri1 29, 1997): "Astoundingly, Germany is one of the few members that fails on debt: its ratio has been rising and is expected to reach 62 per cent at the end of this year. Arguably, other failures are Spain and Austria. Yet Belgium, with a debt ratio of 127 per cent this year, down from 137 per cent in 1993, and Italy, with a ratio of 122 per cent, down from 125 per cent in 1994, can both be judged to meet the criterion. To conclude that Germany has failed and Belgium and Italy have succeeded is palpably absurd. The solution has been simply to ignore the criterion." Back.

Note 11: Some European countries also “borrowed credibility” (from German anti-inflationary policy) during the EMS and subsequently experienced declines in long-term interest rates. See Artus and Bourguinat in Steinherr (1994). Back.

Note 12: For a discussion of the relationship between French domestic politics and the push for EMU, see Milner (1997), Chapter 8. Back.

Note 13: For a brief history of attempts at monetary union in Europe, see Bini-Smaghi et al (1994). Back.

Note 14: In 1989, the UK offered a "competitive currencies" alternative for EMU, in which each member would accept the others' currencies as legal tender (and the "best" currencies would win out). The next year, the British government proposed a "hard ECU," entailing the irrevocable fixing of the ECU in terms of the EC members' currencies, and the circulation of ECUs to the extent that markets demanded them (Grieco, 1993, pp. 271-272; Gros and Thygesen, 1992). Back.

Note 15: Papadia and Saccomanni (in Steinherr, ed., 1994) argue that the difference between gradualism and "shock therapy" represents a false dichotomy. Back.

Note 16: And, ex post, the ERM crises of 1992 probably reinforced this view (Padoa-Schioppa, 1994, pp. 10-12). Back.

Note 17: Holzmann et al (1996) maintain that a more well-defined set of criteria and stronger sanctions for derogation from these criteria are necessary. Gros and Thygesen (1992) propose the creation of a new “EMU indicator,” incorporating inflation, government deficits, government debt, unemployment, and the current account balance (pp. 468-470; also see Artus, 1993). Similarly, Crockett (in Steinherr, 1994) calls for the inclusion of structural variables in the set of convergence criteria. Back.

Note 18: For a similar point regarding domestic scapegoating and the pursuit of anti-inflationary policies during EMS, see Crockett in Steinherr (1994) and Kenen (1995), p. 8. Back.

Note 19: Sandholtz (1993) suggests, however, that without domestic institutional reforms, these gains are limited. “The pro-discipline camp in Italy has scored some important successes, but the credibility of the Italian commitment to discipline remains shaky....Italy's commitment to low inflation will remain suspect as long as political leaders cannot produce the institutional reforms that would separate government spending from the money supply" (p. 9). Back.

Note 20: Governments also addressed the issue of how to define the inflation and interest rate criteria. For critiques of these criteria, see Kenen (1995), pp. 127-132. Back.

Note 21: Gross debt excludes some future government liabilities, especially pension outlays. For arguments that net, rather than nominal gross debt, should be used, see Buiter et al. (1993), pp. 73-74 and Masera in Steinherr (1994), p. 281. Back.

Note 22: This is even more problematic after EMU, where “national fiscal policy activism is in fact the only possible means of adjustment in the short term.” (Artis, 1992, p. 306) Back.

Note 23: Buiter et al (1993) argue that the chosen fiscal criteria are “badly motivated, poorly designed, and apt to lead to unnecessary hardship if pursued mechanically." In a similar vein, Letiche (1993) argues that reducing deficits drastically, in an effort to meet the criteria and deadlines, can lead to or exacerbate economic slowdowns. Back.

Note 24: The “Golden Rule” does have a place in the Treaty, as part of the excessive deficit procedure. “If a Member State does not fulfill the requirements under one or both of these criteria, the Commission shall prepare a report. The report of the Commission shall also take into account whether the government deficit exceeds government investing expenditure and take into account all other relevant factors..." (Treaty on European Union, Article 104c, part 3). Back.

Note 25: Measurement also is a problem with respect to overall levels of debt and deficit: Gros and Thygesen (1992) note that OECD and European Commission statistics for deficit/GDP ratios and debt/GDP ratios vary quite significantly (pp. 280-284). Back.

Note 26: Holzmann et al (1996) criticize the assumption of five percent nominal GDP growth. With a lower level of growth, a three percent deficit/GDP level will not stabilize the debt at or below 60 percent of GDP (p. 40). Also see Crowley (1996) for a similar argument. Back.

Note 27: In their discussion of the convergence criteria, Corsetti and Roubini (in Torres and Giavazzi, 1993) point to the need for an explanation of the reference values, but put aside that issue: “Nonetheless, it would be more satisfactory if some theoretical justification could be given or historical arguments used to support the choice of these numbers, as short run strict guidelines for national fiscal policy. In this section, we by-pass this set of issues and concentrate on a very simple, but important, exercise" (p. 51). Back.

Note 28: See Buiter et al (1993) for a discussion of why, even with a “Golden Rule” rationale, the choice of numbers is economically flawed (p. 63). Back.

Note 29: Additionally, many market participants are rather sanguine regarding the possibility for creative accounting. They note that creative accounting or one-time measures are possible only after a government is near the "target" level. Transferring shares of French Telecom or selling off Italian government industrial holdings is only effective once the French and Italian governments are nearing the three percent deficit level (Interview 35; Interview, European Commission Official, DG-II). Back.