This paper focuses on selected aspects of Italy’s fiscal situation, its origins, and its pervasive economic consequences. The paper analyzes different aspects of regional duality and the unbalanced recovery. The paper highlights that regional duality is a salient feature of the Italian economy. In the south, average incomes and consumption are lower, unemployment rates higher, and the percentage of the population living in poverty larger than in the center and north. The paper also examines fiscal performance of Italy during 1986–95.

Abstract

This paper focuses on selected aspects of Italy’s fiscal situation, its origins, and its pervasive economic consequences. The paper analyzes different aspects of regional duality and the unbalanced recovery. The paper highlights that regional duality is a salient feature of the Italian economy. In the south, average incomes and consumption are lower, unemployment rates higher, and the percentage of the population living in poverty larger than in the center and north. The paper also examines fiscal performance of Italy during 1986–95.

VI. Monetary Policy with High Public Debt 1/

1. Introduction

How does the existence of a large public debt constrain monetary policy? This question is of potential concern in Italy, as well as in several other European countries whose public debts are large in relation to their GDP (Chart 1). In many cases, the fiscal policies leading to large public debts have often been accompanied by relatively high inflation rates compared with those in main partner countries (Chart 2). One interpretation is that public debt limits the central bank’s ability to achieve price stability, making it costly, or even futile, to pursue this goal by monetary policy alone. 2/ If such were the case, disinflation would require the cooperation of fiscal policy. 3/ The issue is not whether such cooperation would be desirable--which it clearly is--but whether, in its absence, monetary policy could or should nonetheless independently pursue price stability.

CHART 1
CHART 1

Debt–to–GDP Ratios, Selected Countries

(In Percent)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

CHART 2
CHART 2

Inflation, Selected Countries

(In percent)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

The link between public debt and inflation is most direct in the absence of central bank independence, where the inflation tax--either anticipated or unanticipated--is another in the array of taxes available to finance government spending to service the debt. 4/ Another instance in which monetary and fiscal policy must clearly march in lock-step is that in which the government has no access to non-monetary financing, or where this access is subject to exogenous quantitative limits. In this case, monetary policy clearly cannot be independent unless it is actually hegemonic--i.e., unless the central bank can dictate the fiscal balance by choosing the financing it is willing to provide.

It is less clear that there should be a link between public debt and inflation in a country with an independent central bank, and where the government has full access to liberalized financial markets. Central bank independence is presumed to free the central bank from the obligation of bending monetary policy to fiscal goals; instead, the central bank should dedicate monetary policy single-mindedly to the goal of price stability. 5/

A counterexample is the case of unpleasant monetarist arithmetic presented by Thomas Sargent and Neil Wallace (1981). The case they examine is one in which the government is pursuing a path of government spending that can be financed only with some use of the inflation tax, at least in the long run. Monetary stringency worsens the debt dynamics as debt is rolled over at higher interest rates, increasing the stock of debt that would need to be monetized later to reach a sustainable path--and thus entailing higher inflation later. Furthermore, if the expectation of higher inflation affects inflation in the current period, a monetary contraction may have a perverse effect, actually increasing current inflation.

The policy implications of unpleasant monetarist arithmetic are clear: inflation is essentially a fiscal phenomenon, and monetary tightening can, at best, buy lower inflation today at the expense of higher inflation tomorrow, but may even lead to higher inflation in all periods. All that the monetary authorities can do is to recognize that an inflation tax is an essential part of the fiscal equation and seek to smooth the required inflation tax over time. This implies that central bank independence is nugatory; disinflation requires action on the fiscal front. 1/

This paper will present some evidence on the possible relevance of unpleasant monetarist arithmetic to two European countries with fiscal imbalances: Ireland, and Italy. The analysis hinges on the interaction between monetary and fiscal policy. Given the complexities of this interaction, it will be important to examine evidence pertaining to the behavior of both monetary and fiscal authorities. To anticipate the results, they suggest that, contrary to the assumptions of unpleasant monetarist arithmetic, fiscal policy does respond significantly to monetary policy and the level of public debt.

The remainder of the paper is structured as follows. Section 2 discusses the concept of unpleasant monetarist arithmetic, emphasizing the underlying assumptions regarding the determinants and interaction of monetary and fiscal policy. Section 3 presents some preliminary empirical evidence: first, a set of vector autoregressions (VARs) examining the interaction of monetary and fiscal policy variables for Italy and Ireland; second, the estimation of cointegrating relationships among these same variables for Italy, having the interpretation of reaction functions; and third, a VAR examining the interaction of money market and forward interest rates. Section 4 presents some tentative conclusions.

2. Unpleasant monetarist arithmetic

Unpleasant monetarist arithmetic implies that even if monetary policy decisions are made separately from fiscal policy decisions, the central bank is constrained by the need to prevent an unsustainable growth of public debt. At the center of the analysis is the government budget constraint:

Dt+Bt(1+it1)Bt1+MtMt1=0(1)

where Dt is the (nominal) primary fiscal balance at time t, Bt the stock of government bonds outstanding at nominal interest rate it and Mt the money stock. Thus, the government can finance a primary deficit either by borrowing or by money creation.

A second key condition is the government solvency condition

limt>βtBt/Pyt=0(2)

where β is the discount factor and Pyt is nominal national income.

These conditions are central to deriving the key results of Sargent and Wallace (1981). They first assume a “monetarist” model in which velocity of money is constant. They adopt a simplified version of condition (2), stating that the debt ratio must be constant from some period T onward. Then they examine the implications of a monetarist rule fixing the rate of growth of the nominal money stock at zero. They show that, for a given path of the primary fiscal deficit Dt, Dt+j,…, this policy necessarily leads to higher inflation from period T onward.

Sargent and Wallace then proceed to extend the simple monetarist model in an obvious direction: introducing the (Bresciani-Turroni-Cagan) effect of expected inflation on money demand. Through this effect, a policy that will generate higher inflation in the future leads, under rational expectations, to higher expected inflation and thus lower demand for money now. Under some conditions, including with regard to the interest elasticity of demand for money, this effect dominates. A perverse result then emerges: contractionary monetary policy may actually be inflationary.

A key element in the argument is the assumption that monetary policy adjusts to accommodate fiscal policy. One possibility of how this could occur is that unsustainable public debt dynamics would, in time, lead to a financial crisis triggering a regime switch that would precipitate a discrete monetization of debt. 1/ In turn, the possibility of this eventuality would be reflected in market interest rates prior to the crisis. Thus, monetary policy could enjoy some freedom of maneuver for some period of time but periodically be forced into line with the requirements of the fiscal situation--and the awareness of this limitation to central bank independence may be built into interest rates now.

This analysis, which characterizes inflation as a purely fiscal phenomenon, must be qualified on several grounds. First, it is predicated on the assumption that fiscal policy would be unsustainable unless the monetary authorities play an equilibrating role--leaving no degrees of freedom for monetary policy. This requires that the real interest rate on government debt exceed the real growth rate of the economy--an assumption that has been challenged in the context of the U.S.. 1/ In several high-public-debt European countries, however, real interest rates have generally exceeded real growth rates (Chart 3), suggesting that unpleasant monetarist arithmetic may be of greater relevance there.

CHART 3
CHART 3

Real Interest Rates and Growth Rates, Selected Countries

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

Source: World Economic Outlook and IMF, International Financial Statistics.1/ For 1995, average of January-November.

A second qualification of the Sargent-Wallace argument is that it hinges on a particular definition of “a given fiscal policy.” A constant overall fiscal deficit--i.e., including debt servicing--is not inflationary if it is bond-financed, but a constant bond-financed primary deficit is inflationary (McCallum (1984)). Thus unpleasant monetarist arithmetic would not hold if the primary fiscal balance is adjusted to maintain a given overall balance in the face of variations in debt interest costs. 2/

More generally, unpleasant monetarist arithmetic assumes a particular structure of relations between government and central bank. Although the central bank acts independently, it is assumed to take the government’s actions as given--that is, to leave the government the role of Stackelberg leadership. The central bank is left in the position of ensuring the sustainability of an otherwise unsustainable fiscal policy. One can also visualize the reverse case--which could be characterized as market discipline--in which the central bank credibly sets a course of monetary policy independently of the government’s actions, and fiscal policy must adjust to ensure the government’s solvency. In this case, an unsustainable fiscal policy would result in increasing interest rates--to the extent that it resulted in increasing prospects of bankruptcy--in turn providing an incentive for the government to adjust to a sustainable path of fiscal policy (Lane 1993). In that case, monetary tightening would put pressure for primary fiscal adjustment.

In examining the implications of high public debt for monetary policy, it is therefore essential to consider the reactions of both the government and the central bank. However, institutional arrangements may affect the rules of the game between the two. For Instance, with “the divorce” in 1981, the Bank of Italy was no longer required to accommodate the government’s debt financing--and Tabellini (1988) found that the degree of monetary accommodation actually declined. Exchange rate arrangements may also alter the degree of monetary accommodation that need take place: for instance, joining a monetary union could, in principle, remove the possibility of monetizing public debt, requiring that all of the adjustment is on the fiscal side.

3. Empirical results

This section presents some empirical evidence on unpleasant monetarist arithmetic for two high-debt European countries, Ireland and Italy. Further work should be done to extend the analysis to other countries and examine the robustness of the results with respect to alternative specifications. However, the results presented here provide some indications of the interaction of monetary and fiscal policy that is central to this discussion.

a. Monetary and fiscal policy interaction

This section discusses the results of vector autoregressions examining the links between monetary and fiscal policy. Data are quarterly for 1979-1995. 1/ For Ireland, all variables are stationary. For Italy, some are non-stationary, with two cointegrating vectors; these cointegrating vectors will be examined in the next subsection. 2/

Vector autoregressions are estimated using the following variables: general government gross public debt, the discount rate, the inflation rate, GDP, and the primary fiscal balance. 3/

The charts show the responses to a 1-standard error shock to end of the relevant variables. One noteworthy result is that a shock to the discount rate has a negative overall effect on inflation, with a similar response in the two countries (Chart 4)--although for Ireland the response is somewhat stronger and more persistent. This is contrary to unpleasant monetarist arithmetic, suggesting that over the sample period monetary policy was effective in both countries, notwithstanding their large public debts. 4/

CHART 4
CHART 4

Impulse Response of Inflation to Discount Rate Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

Shocks to the debt stock had a negative effect on inflation in both countries (Chart 5). Under unpleasant monetarist arithmetic, an Increase in debt, other things equal, would be inflationary, as it would imply more debt to be monetized now. The empirical results here are more supportive of the more traditional view that if more debt is monetized (implying slower growth of debt for given primary balance, past debt, and interest rates) this results in higher inflation.

CHART 5
CHART 5

Impulse Response of Inflation to Debt Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

The portion of the variation in inflation explained by various factors can be seen in the variance decompositions shown in Chart 6: this chart makes it clear that inflation is more strongly affected by the discount rate in Ireland than in Italy. The debt stock explains close to a third of the variance in inflation in Ireland, although much less in Italy. The primary fiscal balance has a small effect in Italy but a negligible one in Ireland.

CHART 6
CHART 6

Forecast Error Variance Decomposition of Inflation

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

Another key result concerns the determinants of the primary fiscal balance. An increase in the discount rate has a positive effect on the primary fiscal balance in both countries (Chart 7). In Italy, however, this effect is snail In magnitude and significant at the 10 percent level only for quite a long time horizon. For Ireland, the effect is much more uniformly positive and significant at a shorter lag. The effect of the discount rate to the primary balance may be transmitted through the stock of public debt; the discount rate has a positive effect on the debt stock in both countries (Chart 8), which in Ireland tapers off while in Italy it persists (although it is not amplified over time, as suggested by unpleasant monetarist arithmetic). At the same time, the debt stock has a significant effect on the primary fiscal balance in both countries--although again, in Ireland this effect is much stronger than in Italy (Chart 9). 1/

CHART 7
CHART 7

Impulse Response of Primary Balance to Discount Rate Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

CHART 8
CHART 8

Impulse Response of Debt-to-GDP Ratio to Discount Rate Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

CHART 9
CHART 9

Impulse Response of Primary Balance to Debt Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

These results lend some support for the market discipline view that higher interest rates spur primary adjustment. 2/ This is also consistent with McCallum’s (1984) condition for debt-financed deficits to be non-inflationary--that an increase in debt servicing costs bring about primary adjustment which tends to stabilize the overall balance. However, whereas for Ireland a shock to the interest rate eventually stimulates enough primary adjustment to bring the stock of debt back to its initial level, for Italy a shock to the discount rate brings about a persistent increase in the debt stock. This suggests that the effect of monetary policy on the debt dynamics is likely to be of greater concern in Italy than in Ireland. 3/

Chart 11 shows the response of the discount rate to the shock to inflation: in both countries, the authorities responded to unexpectedly high inflation by raising the discount rate, but the response was quantitatively larger and prompter in Ireland than in Italy. This suggests that the Central Bank of Ireland pursued a more aggressively anti-inflationary policy over this period than the Bank of Italy.

CHART 10
CHART 10

Impulse Response of Debt-to-GDP Ratio to Primary Balance Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

CHART 11
CHART 11

Impulse Response of Discount Rate to Inflation Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

Finally, one can examine the effects of inflation shocks on fiscal policy, as shown in the response of the primary balance to an inflation shock (Chart 12). In the case of Italy, there is a significantly positive effect, possibly reflecting fiscal drag. In Ireland, the effect is approximately zero.

CHART 12
CHART 12

Impulse Response of Primary Balance to Inflation Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

The variance decomposition of the primary balance shown in Chart 13 sheds further light on the issue. This confirms the small effect of inflation on the primary balance In Ireland, as compared with Italy. More importantly, it suggests that while fiscal policy In both Ireland and Italy is somewhat responsive to accumulated debt and to the discount rate, that responsiveness is substantially greater in Ireland than in Italy. This is not directly related to the relative size of the two countries’ debt, as Ireland’s exceeded Italy’s until 1989. It may instead reflect some aspect of the two countries’ institutional structure or political processes which have made Italian fiscal policymakers more oblivious to the mounting costs of financing the public debt.

CHART 13
CHART 13

Forecast Error Variance Decomposition of Primary Balance

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

b. Reaction function?

A second approach that may shed light on the interaction of monetary and fiscal policy with high public debt is the estimation of reaction functions. For this purpose, the same data set as in the previous section is used, focussing on Italy. For Italy, the variables in the data set--the debt-to-GDP ratio, the primary fiscal balance as a percent of GDP, the discount rate, and the inflation rate--were found to be non-stationary over the sample period. The hypothesis of two cointegrating vectors among these variables could not be rejected. These cointegrating vectors can be represented as follows:

Dt=2.57rtd+0.510Bt(4)

and

rtd=0.397πt0.683Bt(5)

Equation (4) can be interpreted as a fiscal policy reaction function. It suggests that the primary balance is strengthened when the government is faced with higher interest rates, and the primary balance adjusts to offset the accumulation of public debt. This is broadly supportive of the market discipline story, as opposed to the unpleasant monetarist arithmetic story. The estimated magnitudes of the reactions are surprisingly large.

Equation (5) can be interpreted as a monetary policy reaction function. It suggests that the central bank sets higher official interest rates when inflation is higher. 1/ Secondly, the analysis suggests that the central bank does not fully accommodate the effects of fiscal policy but, on the contrary, has shown a tendency to raise interest rate as the primary balance weakens. This estimated reaction function is therefore also inconsistent with unpleasant monetarist arithmetic.

Caution is clearly required in interpreting these cointegration results, as well as the VAR estimates presented in the previous section. They reflect the behavior of monetary and fiscal policy during a particular period, and may not be robust across sample periods. In particular, they do not address the issue of whether Italian monetary policy may have become more of a hostage to fiscal policy with the rising public debt. Moreover, within the ERM, Italian monetary policy was largely tied to conditions elsewhere in Europe, and this may have helped stiffen the Bank of Italy’s hand against the demands of the Treasury; in this light, the lira’s departure from the ERM may have reduced the independence of Italian monetary policy. However, there is also the opposite possibility: that with the deepening of financial markets and the formal independence of the Bank of Italy, the central bank’s capacity for autonomous action to quell inflation is now larger than before. The post-ERM sample period is unfortunately too short to discriminate between these alternatives.

c. Money market and forward interest rates

Some further indications of the relevance of unpleasant monetarist arithmetic can be gleaned from a VAR involving overnight interest rates, forward interest rates, and inflation. 1/ Unpleasant monetarist arithmetic implies that if a monetary tightening pushes up overnight interest rates, this should worsen the debt dynamics, resulting in higher expected inflation, as reflected in forward interest rates, and thus higher actual inflation. 2/ The traditional story has the reverse implication: that monetary tightening lowers expected inflation, and thus lowers forward interest rates, as well as lowering actual inflation.

Monthly data on forward interest rates, from the swap market, were examined for the post-ERM period December 1992 through September 1995. The five-year forward interest rate was used. Other variables in the VAR were overnight interest rates and consumer price inflation. The most important result is with regard to the response of the forward rate to a shock to overnight rates (Chart 14): the initial impact is positive in the very short run--possibly reflecting bond market liquidity effects (Angeloni and Prati (1993)), while in the long run it turns negative; the effect is, however, insignificant over the whole horizon. The evidence thus fails to support the unpleasant monetarist arithmetic story, but also it provides at best very weak support for the alternative, monetarist viewpoint. The response of inflation to a shock to money market interest rates is also insignificant.

Chart 14
Chart 14

Italy. Impulse Response of 5-year Forward Rate to Overnight Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

Another interesting result, not directly related to unpleasant monetarist arithmetic, is the response of the overnight rate to a shock to forward rates (Chart 15). This effect is positive and persistent, suggesting that over this period the Bank of Italy “followed the market” in influencing money market interest rates.

CHART 15
CHART 15

Italy. Impulse Response of Overnight to 5-year Forward Rate Shock

(size of confidence bounds = 10%)

Citation: IMF Staff Country Reports 1996, 035; 10.5089/9781451819694.002.A006

4. Conclusion

This paper has presented some preliminary evidence that sheds some light on the relevance of unpleasant monetarist arithmetic to two high-debt European countries, Italy and Ireland. The results tend to suggest that fiscal policy is more responsive to monetary tightening, and to the accumulation of public debt, than the unpleasant monetarist arithmetic hypothesis would allow; these effects appear to be stronger for Ireland than for Italy, despite the fact that Ireland had a higher debt-to-GDP ratio for much of the sample period. Monetary policy does not appear to have fully accommodated variations in the fiscal balance. Moreover, the evidence suggests that monetary policy has been somewhat effective in influencing inflation, for a given primary fiscal balance and debt--again, with stronger effect for Ireland than for Italy. Estimated cointegrating equations for Italy, which can be Interpreted as reaction functions for the central bank and government, also suggest that monetary policy has not been fully accommodative, nor has fiscal policy remained impervious to market discipline.

It should be emphasized that these results are very tentative: they require further investigation of their robustness with respect to sample period and variables included. Additional tests would also help strengthen or qualify the conclusions in the paper.

The results so far provide little support for the view that a high public debt should deter the central bank from pursuing the goal of price stability. The evidence presented cannot rule out the possibility that monetary policy may have become less effective in recent years, as the stock of public debt has grown. Moreover, participation in the ERM for much of the sample period may have tied the central bank’s hand, dictating that at least some accommodation come from the fiscal side; the lira’s departure from the ERM may therefore have heralded a monetary policy setting that was more conducive to accommodation. However, at the same time, other developments may have reduced the ability of fiscal policy makers to ignore the need to finance the public debt; in Italy, these considerations have included the abolition of capital controls, the heightened sensitivity of bond markets to news, and the 1993 law establishing the formal independence of the Bank of Italy (including by prohibiting monetary financing). Such changes may have eased the fiscal constraints on monetary policy.

References

Chapter VI

  • Aiyagari, S. Rao, and Mark Gertler (1995) “The Backing of Government Bonds and Monetarism,” Journal of Monetary Economics Vol. 16, No. 1 (July), pp. 19-44.

    • Search Google Scholar
    • Export Citation
  • Angeloni, Ignazio, and Alessandro Prati (1993), “Liquidity Effects and the Determinants of Short-Term Interest Rates in Italy (1991-92)” Center for Economic Policy Research Discussion Paper 788, May, pp. 1-22.

    • Search Google Scholar
    • Export Citation
  • Buiter, Willem (1982), “Comment on T.J. Sargent and N. Wallace, ‘Some Unpleasant Monetarist Arithmetic’,” NBER Working Paper 867, March.

    • Search Google Scholar
    • Export Citation
  • Calvo, Guillermo (1978), “On the Time Consistency of Optimal Policy in a Monetary Economy,” Econometrica Vol. 46: pp. 1411-1428.

  • Carmichael, Jeffrey, and Peter W. Stebbing (1983), “Fisher’s Paradox and the Theory of Interest,” American Economic Review Vol. 73, No. 4 (September), pp. 619-630.

    • Search Google Scholar
    • Export Citation
  • Darby, Michael R. (1984), “Some Pleasant Monetarist Arithmetic,” Federal Reserve Bank of Minneapolis Quarterly Review Vol. 8, No. 2 (Spring), pp. 15-20. Reply by Preston J. Miller and Thomas J. Sargent, pp. 21-26.

    • Search Google Scholar
    • Export Citation
  • Drudi, Francesco, and Alessandro Prati (1995), “Signalling Fiscal Regime Sustainability”. Mimeo, Bank of Italy and IMF, November.

  • Giavazzi Francesco, and Marco Pagano (1990), “Confidence Crises and Public Debt Management,” in Public Debt Management in Theory and History edited by Rudiger Dorhbusch and Mario Draghi. Cambridge: Cambridge University Press.

    • Search Google Scholar
    • Export Citation
  • Giavazzi, Francesco, and Luigi Spaventa (1989), “Italy: the Real Effects of Inflation and Disinflation”. Economic Policy No. 4.

  • Kneebone, Ronald D., and William M. Scarth (1990), “Is Non-Monetarist Arithmetic Just As Unpleasant?” Southern Economic Journal 57, No. 1 (July), pp. 14-21.

    • Search Google Scholar
    • Export Citation
  • Lane, Timothy D. (1993), “Market Discipline,” IMF Staff Papers 40, No. 1 (March), pp. 53-88.

  • Lane, Timothy D. (1995), “High Public Debt: Consequences for Investment and Growth”. Paper presented at conference on Banking, Foreign Investment, and Capital Flows, University of Potsdam, October.

    • Search Google Scholar
    • Export Citation
  • Lane, Tinothy D., Alessandro Prati, and Mark E.L. Griffiths (1995), “An Inflation Targeting Framework for Italy,” IMF Paper on Policy Analysis and Assessment, PPAA/95/4, March.

    • Search Google Scholar
    • Export Citation
  • McCallum, Bennett T. (1984), “Are Bond-Financed Deficits Inflationary?” Journal of Political Economy, 92 No. 1 (February), pp. 123-135.

    • Search Google Scholar
    • Export Citation
  • Sargent, Thomas J., and Neil Wallace (1981), “Some Unpleasant Monetarist Arithmetic,” Federal Reserve Bank of Minneapolis Quarterly Review 5, no 3 (Fall), pp. 1-17.

    • Search Google Scholar
    • Export Citation
  • Tabellini, Guido (1988), “Monetary and Fiscal Policy Coordination with a High Public Debt”. In High Public Debt: The Italian Experience, Cambridge U.K.: Cambridge University Press.

    • Search Google Scholar
    • Export Citation
  • Visco, Ignazio (1995), “Inflation, Inflation Targeting, and Monetary Policy. Notes for Discussion on the Italian Experience.” In Inflation Targets edited by Leonardo Leiderman and Lars Svensson. London: CEPR.

    • Search Google Scholar
    • Export Citation
1/

Prepared by Timothy Lane and Alessandro Prati.

2/

There are of course other possible interpretations: for instance, that similar institutional factors have led to weaknesses in both fiscal and monetary policies in some countries, or that supply shocks have affected both inflation and fiscal balances.

3/

The Bank of Italy view that successful disinflation requires a “triangle” of policies--monetary, fiscal, and incomes--is discussed in Visco (1995). Clearly, this does not hinge on the unpleasant monetarist arithmetic discussed later in this paper, but only on the premise that disinflation may be unduly costly if pursued by monetary policy alone.

4/

As discussed in the literature on time consistency, a policy of surprise inflation is futile: since the incentive to levy a tax on public debt through surprise inflation is itself predictable, the inflation comes as no surprise. However, this temptation to surprise the public may nonetheless result in a higher fully-anticipated inflation rate. See, for instance, Calvo (1978).

5/

For instance, this argument is made in the Italian context by Lane, Griffiths and Prati, 1995.

1/

A related idea is the “backing” of government bonds (Aiyagari and Gertler (1985)): the fraction of the public debt that will be serviced through future taxes--where the rest corresponds to future money creation. If bonds are fully backed, inflation is a monetary phenomenon but, if not, public debt is inflationary for a given current monetary policy.

1/

Giavazzi and Pagano (1990), for instance, analyzed such a confidence crisis under a fixed-exchange-rate regime: an attack on the currency which would face the central bank with the choice of accepting a currency depreciation and inflation associated with debt monetization, or raising interest rates to defend the currency at the cost of an explosion of public debt. The central bank’s incentive to give way then depends on the size and structure of public debt.

1/

See Darby (1984). Sargent’s reply (co-authored with Preston Miller) argued that the real interest rate could change as the debt-to-GDP ratio increases, reasserting the potential relevance of unpleasant monetarist arithmetic to the U.S.

2/

These results would also be altered if high public debt had real effects on investment and growth. An analytical overview of potential effects is presented in Lane (1995).

1/

In this version, quarterly fiscal data are interpolated from annual data. Further work will include the use of quarterly fiscal data for Italy from the Relzione Trimestrale di Cassa.

2/

Due to the small sample size, current and one lagged value of each variable were used; this lag length was accepted by the Akaike and Schwarz information criteria.

3/

Another potential variable for inclusion is the exchange rate. However, since during this period the exchange rate was constrained by the ERM bands, it is unlikely to have displayed enough variation to be empirically relevant. This is an obvious area for further work.

4/

A possible caveat is that inflation over this period may have been strongly influenced by the oil price shocks. In preliminary work (not reported here), it was found that the results were not substantially altered by including the terms of trade in the VAR, although all the results became weaker as would be expected by adding an additional variable.

1/

Surprisingly, however, despite the rules of accounting, the primary balance did not have a significant negative effect on the debt stock in either country (Chart 10).

2/

Part of the effect may have been the mechanical effect of higher interest income on revenue from the withholding tax. This is probably a minor part of the overall effect, but warrants further investigation.

3/

Similar empirical results are interpreted in light of a signalling model in Drudi and Prati (1995). In their model, primary budget deficits lead to a buildup of public debt, which in turn generates a separating equilibrium in which governments signal their fiscal probity through primary fiscal adjustment.

1/

Although the magnitude of the coefficient is smaller than predicted by the Fisher effect, such a result is not unusual from simple regression analyses (e.g., Carmichael and St ebbing, 1983).

1/

One topic for future exploration is the possibility that the exchange rate should also be included in this VAR--although the volatility of the exchange rate over this sample period, apparently in response to political factors, suggests that it is an unlikely candidate.

2/

Strictly speaking, the Sargent and Wallace (1981) model, which ignores any liquidity effects, implies that in the case of unpleasant monetarist arithmetic a monetary tightening also lowers overnight interest rates.

Italy: Background Economic Issues
Author: International Monetary Fund
  • View in gallery

    Debt–to–GDP Ratios, Selected Countries

    (In Percent)

  • View in gallery

    Inflation, Selected Countries

    (In percent)

  • View in gallery

    Real Interest Rates and Growth Rates, Selected Countries

  • View in gallery

    Impulse Response of Inflation to Discount Rate Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Impulse Response of Inflation to Debt Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Forecast Error Variance Decomposition of Inflation

  • View in gallery

    Impulse Response of Primary Balance to Discount Rate Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Impulse Response of Debt-to-GDP Ratio to Discount Rate Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Impulse Response of Primary Balance to Debt Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Impulse Response of Debt-to-GDP Ratio to Primary Balance Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Impulse Response of Discount Rate to Inflation Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Impulse Response of Primary Balance to Inflation Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Forecast Error Variance Decomposition of Primary Balance

  • View in gallery

    Italy. Impulse Response of 5-year Forward Rate to Overnight Shock

    (size of confidence bounds = 10%)

  • View in gallery

    Italy. Impulse Response of Overnight to 5-year Forward Rate Shock

    (size of confidence bounds = 10%)