Italy: Basic Data

Area: (Thousands sq. km) 301.336 Population: (millions, 1998) 57.6 GDP per capita: 1999 20,356 (Percentage changes, except as otherwise indicated)

Sources: Data provided by the Italian authorities; and Fund staff estimates and projections.

^1/Staff estimates and projections, unless otherwise indicated.

^2/Volumes and unit values are customs basis; trade balance and current account are balance of payments basis.

^3/Data for 2000 refer to first quarter.

^4/Data for 2000 refer to January.

^5/End-of-period; data break in 1998.

^6/Data for 2000 refer to February.

^7/Period average.

Area: (Thousands sq. km) 301.336 Population: (millions, 1998) 57.6 GDP per capita: 1999 20,356 (Percentage changes, except as otherwise indicated)

			1995	1996	1997	1998	1999	2000 1/
Domestic economy
GDP			2.9	1.1	1.8	1.5	1.4	2.7
Domestic demand			2.0	0.9	2.5	2.9	2.5	2.5
	Consumption (households)		1.7	1.2	3.0	2.3	1.7	1.8
	Public consumption		−2.1	1.1	0.9	0.7	0.8	0.2
	Gross fixed investment		6.0	3.6	1.2	4.1	4.4	6.4
		Machinery and equipment	10.6	3.7	4.2	7.4	6.2	7.5
		Construction	0.9	3.6	−2.3	−0.1	1.8	4.7
	Inventories (contribution)		0.2	−0.7	0.3	0.6	0.4	0.1
Foreign sector (contribution)			1.0	0.2	−0.6	−1.3	−1.0	0.3
	Exports		12.6	0.6	6.5	3.3	−0.4	8.0
	Imports		9.7	−0.3	10.2	9.1	3.4	7.4
Employment			−0.6	0.5	0.4	1.1	1.3	1.1
Unemployment			11.6	11.6	11.7	11.8	11.4	11.0
Manufacturing value added			4.7	−1.3	2.3	1.9	1.2	3.5
Labor costs			4.7	5.8	4.6	−2.0	2.4	2.5
Unit labor costs in manufacturing			1.1	6.1	2.2	−1.8	0.9	0.4
Consumer prices (period average)			5.2	3.9	1.7	1.8	1.7	2.2
GDP deflator			5.0	5.3	2.4	2.7	1.5	1.9
Output gap			−1.1	−2.0	−2.3	−2.7	−3.1	−2.5
External accounts 2/
Export volume			13.2	−2.2	5.6	3.9	−1.2	8.0
Import volume			9.7	−4.3	11.2	11.1	0.4	7.4
Export unit value			9.3	4.3	−0.3	0.3	−0.4	5.0
Import unit value			12.3	0.0	0.1	−4.7	3.7	7.1
Trade balance (in percent of GDP)			3.5	4.4	3.4	3.1	1.8	1.6
Current account (in percent of GDP)			2.3	3.2	2.8	1.8	0.9	0.8
Nominal effective exchange rate 3/			−9.8	9.3	0.7	−0.4	−1.8	−3.4
Real effective exchange rate 3/			−7.5	11.9	2.8	0.5	−2.0	−3.7
Public finances (in percent of GDP)
General government
Revenues			45.6	45.8	48.2	46.6	46.9	46.6
Expenditures			53.2	52.9	50.9	49.4	48.8	48.1
Balance			−7.6	−7.1	−2.7	−2.8	−1.9	−1.5
Primary balance			3.9	4.4	6.7	5.3	4.9	5.1
Structural primary balance			4.5	5.3	7.9	6.6	6.5	6.3
Debt			123.2	122.2	119.8	116.3	114.9	110.7
State sector
	Balance		−7.1	−6.8	−2.6	−2.8	−1.5	−1.0
Financial variables
M3 4/			…	…	…	1.6	3.4	4.3
Total domestic credit 5/ 6/			4.3	5.2	3.5	4.4	7.8	6.8
	Of which: To nonstate sector 5/ 6/		2.4	3.2	5.2	9.2	13.4	12.9
Six-month rate on Treasury bills 3/ 7/			10.9	8.6	6.4	5.0	3.0	3.8
Prime lending rate 4/ 7/			11.0	11.0	9.2	7.7	5.9	6.3

Sources: Data provided by the Italian authorities; and Fund staff estimates and projections.

^1/Staff estimates and projections, unless otherwise indicated.

^2/Volumes and unit values are customs basis; trade balance and current account are balance of payments basis.

^3/Data for 2000 refer to first quarter.

^4/Data for 2000 refer to January.

^5/End-of-period; data break in 1998.

^6/Data for 2000 refer to February.

^7/Period average.

View Table

Area: (Thousands sq. km) 301.336 Population: (millions, 1998) 57.6 GDP per capita: 1999 20,356 (Percentage changes, except as otherwise indicated)

			1995	1996	1997	1998	1999	2000 1/
Domestic economy
GDP			2.9	1.1	1.8	1.5	1.4	2.7
Domestic demand			2.0	0.9	2.5	2.9	2.5	2.5
	Consumption (households)		1.7	1.2	3.0	2.3	1.7	1.8
	Public consumption		−2.1	1.1	0.9	0.7	0.8	0.2
	Gross fixed investment		6.0	3.6	1.2	4.1	4.4	6.4
		Machinery and equipment	10.6	3.7	4.2	7.4	6.2	7.5
		Construction	0.9	3.6	−2.3	−0.1	1.8	4.7
	Inventories (contribution)		0.2	−0.7	0.3	0.6	0.4	0.1
Foreign sector (contribution)			1.0	0.2	−0.6	−1.3	−1.0	0.3
	Exports		12.6	0.6	6.5	3.3	−0.4	8.0
	Imports		9.7	−0.3	10.2	9.1	3.4	7.4
Employment			−0.6	0.5	0.4	1.1	1.3	1.1
Unemployment			11.6	11.6	11.7	11.8	11.4	11.0
Manufacturing value added			4.7	−1.3	2.3	1.9	1.2	3.5
Labor costs			4.7	5.8	4.6	−2.0	2.4	2.5
Unit labor costs in manufacturing			1.1	6.1	2.2	−1.8	0.9	0.4
Consumer prices (period average)			5.2	3.9	1.7	1.8	1.7	2.2
GDP deflator			5.0	5.3	2.4	2.7	1.5	1.9
Output gap			−1.1	−2.0	−2.3	−2.7	−3.1	−2.5
External accounts 2/
Export volume			13.2	−2.2	5.6	3.9	−1.2	8.0
Import volume			9.7	−4.3	11.2	11.1	0.4	7.4
Export unit value			9.3	4.3	−0.3	0.3	−0.4	5.0
Import unit value			12.3	0.0	0.1	−4.7	3.7	7.1
Trade balance (in percent of GDP)			3.5	4.4	3.4	3.1	1.8	1.6
Current account (in percent of GDP)			2.3	3.2	2.8	1.8	0.9	0.8
Nominal effective exchange rate 3/			−9.8	9.3	0.7	−0.4	−1.8	−3.4
Real effective exchange rate 3/			−7.5	11.9	2.8	0.5	−2.0	−3.7
Public finances (in percent of GDP)
General government
Revenues			45.6	45.8	48.2	46.6	46.9	46.6
Expenditures			53.2	52.9	50.9	49.4	48.8	48.1
Balance			−7.6	−7.1	−2.7	−2.8	−1.9	−1.5
Primary balance			3.9	4.4	6.7	5.3	4.9	5.1
Structural primary balance			4.5	5.3	7.9	6.6	6.5	6.3
Debt			123.2	122.2	119.8	116.3	114.9	110.7
State sector
	Balance		−7.1	−6.8	−2.6	−2.8	−1.5	−1.0
Financial variables
M3 4/			…	…	…	1.6	3.4	4.3
Total domestic credit 5/ 6/			4.3	5.2	3.5	4.4	7.8	6.8
	Of which: To nonstate sector 5/ 6/		2.4	3.2	5.2	9.2	13.4	12.9
Six-month rate on Treasury bills 3/ 7/			10.9	8.6	6.4	5.0	3.0	3.8
Prime lending rate 4/ 7/			11.0	11.0	9.2	7.7	5.9	6.3

Sources: Data provided by the Italian authorities; and Fund staff estimates and projections.

^1/Staff estimates and projections, unless otherwise indicated.

^2/Volumes and unit values are customs basis; trade balance and current account are balance of payments basis.

^3/Data for 2000 refer to first quarter.

^4/Data for 2000 refer to January.

^5/End-of-period; data break in 1998.

^6/Data for 2000 refer to February.

^7/Period average.

Introduction and Overview

1. Following Italy’s successful drive to achieve founder membership in European Economic and Monetary Union (EMU), the focus of the economic policy debate is shifting toward structural, longer-term issues. Against the background of a decade of weak economic growth, the key challenge is to identify and implement policies that would strengthen growth, and spread its fruits more widely in terms of employment creation and regional economic performance.

2. The three chapters of this paper shed light on several, closely related, aspects of this challenge—and in particular on the contribution that fiscal policy can make, together with flanking structural reforms. The analytical work reported in Chapter I attempts to identify the sources of Italy’s weak growth performance from a cross-country perspective—and attributes it, at least in part, to a high and rising tax burden. While lowering the tax burden would therefore seem desirable, fiscal policy is severely constrained not only by the Stability and Growth Pact, but also by the still high level of public debt and the future fiscal burden related to population aging. These longer-run fiscal trade-offs, and their implications for growth, are discussed in Chapter II. One fiscal sector that is likely to be affected strongly by population aging is health care, and past reforms and future challenges are discussed in Chapter III; issues related to pensions were analyzed in last year’s background papers (see Chapter I in SM/99/115, 5/20/99). The rest of this section provides a brief overview of the three chapters.

3. Chapter I—Puzzling Out Italy’s Growth Performance—investigates the factors behind Italy’s relatively weak growth during the past decade. Employing a cross-country approach, Italy’s experience is set against that of France, the Netherlands, and the United Kingdom. On the demand side, the empirical results confirm past work, suggesting that restrictive fiscal and monetary policies—indispensable for participation in EMU—weakened growth to some extent. However, these demand-side factors account for only a small part of Italy’s growth differential vis-à-vis the comparator countries—notably, the Netherlands and the United Kingdom—suggesting that structural factors may have played a central role. A simple growth accounting exercise suggests that low employment growth (as opposed to weak capital or total factor productivity growth) was the main culprit behind Italy’s poor growth performance; accordingly, the chapter focuses on structural factors in the labor market, including the role of labor taxation.

4. The labor market analysis decomposes the observed labor market movements into structural labor demand and supply factors. On the labor demand side, there is evidence in most countries of a fall in demand over the past two decades, reflecting exogenous technological shifts and, in Italy especially in the 1990s, perhaps also a reduction of “featherbedding” as union power declined and a major privatization program took hold. More importantly, there is evidence that labor supply in Italy expanded much less than in the comparator countries, as wages (adjusted for productivity) responded to a much smaller degree to unemployment than in the comparator countries. Moreover, the results point to regional asymmetries in Italy, with rising unemployment in the South having no discernible impact on productivity-adjusted wages. In assessing the relatively small response of pretax wages to adverse labor market developments, the analysis suggests that the sharp increase in the fiscal burden in Italy during the past decade may have been an important factor: the squeeze on (after-tax) wages due to rising taxes made pre-tax wage moderation more difficult than, for example, in the Netherlands, where the labor tax burden fell substantially over this period, raising household income.

5. The need for a durable reduction of the fiscal burden is one of the considerations underlying the discussion of an appropriate medium- and long-term fiscal strategy—the topic of Chapter II: Italy’s Fiscal Strategy in a Medium-Term Framework. The analysis is anchored on long-term considerations that take explicitly into account the impact on the public finances of population aging—projected to be more adverse than in most industrial countries. Population aging affects the public finances in mainly two ways: first, it raises public expenditure, notably on pensions and health care (only partially offset by a decline in spending on education); and, second, it lowers the rate of potential output growth, as the labor force declines.

6. Against this background, two alternative fiscal scenarios are discussed: a baseline scenario, which broadly follows the Italian authorities’ Stability Program through 2003 and maintains a balanced budget over the longer-run, with some reductions in the fiscal burden and the ratio of expenditure to GDP; and a more ambitious reform scenario, aiming at faster debt reduction (with the primary surplus maintained at somewhat over 6 percent of GDP over the cycle, resulting in substantial surpluses in the overall balance by the end of the current decade), along with deeper revenue and expenditure cuts. A comparison between the two scenarios illustrates some of the advantages of aiming at relatively fast, early debt reduction. While under the baseline scenario the debt ratio keeps falling monotonically, its pace of decline is not fast enough to bring it below the Stability and Growth Pact ceiling of 60 percent of GDP by the time the most adverse phase of the demographic shock takes place; this leaves very little room to accommodate the shock, and the fiscal burden has to rise precisely at the time when the decline in the labor force is steepest. By contrast, the more ambitious reform scenario attains a debt ratio below the 60 percent reference value much earlier (around 2010). This leaves sufficient room for the deficit and debt ratios to temporarily rise again during the most adverse phase of the demographic shock so as to keep the fiscal burden unchanged. These long-run simulations are subject to considerable uncertainty about the underlying structural parameters, and the chapter also discusses some alternative specifications (for example, with respect to labor productivity growth).

7. The results in Chapters I and II underscore the pivotal role of public expenditure restraint, needed not only for public debt reduction but also for lowering the tax burden and strengthening growth—a politically difficult task as demonstrated by the Italian experience and that of other countries. Chapter III—The Evolving Role of Regions in Italy: The Financing and Management of Health Care Services—focuses on one key expenditure area, health care spending. Health care spending is of particular importance for at least two reasons: it is, aside from pensions, the expenditure category that will be most adversely affected by the upcoming demographic shock; and it constitutes the bulk of regional government expenditure, where sizable overruns have emerged in the past. Moreover, the issue is particularly topical, as the Italian authorities are currently in the midst of reforming the health care system, including through increased fiscal devolution to the regions.

8. The recent and pending reforms of the health care system are raising the regions’ own resources, including through expanded discretion in setting rates on local taxes; and it replaces the old system of central government transfers by an equalization mechanisms that specifies interregional transfers on the basis of regional tax bases (depending on levels of activity) and health care needs (depending on demographic characteristics). The new system—which is also reviewed from a cross-country perspective—improves fiscal transparency by setting clear rules for intergovernmental fiscal relationships. Nonetheless, potential problems could arise if: systematic differences in administrative capacity among regions are not addressed; growth differences persist between regions of a scale experienced in the past; and slow-growth regions are forced to raise their regional tax rates (which could aggravate regional disparities, should the tax base prove sufficiently mobile). Moreover, present reforms do not address the pending pressure on health care expenditure due to population aging—central if the challenges are to be addressed that were identified in Chapters I and II, and a point that is well recognized by the authorities.

I. Puzzling Out Italy’s Growth Performance¹

9. Economic growth in Italy has been subdued during the 1990s, which, in turn, has prompted a search for explanations. With the objective of ultimately reaching policy conclusions, this chapter evaluates the merits of various hypotheses that have been advanced to explain Italy’s performance. Since both the decline in growth by comparison with the 1980s and Italy’s underperformance relative to European Union (EU) partner countries as well as most other industrial countries have attracted attention, the paper takes a cross-country approach in assessing the various hypotheses. The comparator countries chosen are France, the Netherlands, and the United Kingdom. Focusing on several countries rather than on EU averages, for example, makes it easier to pinpoint structural or demand-side explanatory factors for differing growth rates; these factors are easily concealed by averages. Moreover, the three countries are fairly heterogeneous in their macroeconomic and structural policies as well as growth outcomes, thereby fostering more interesting comparisons with Italy: France, together with Italy, recorded below EU-average growth, while the United Kingdom and the Netherlands benefited, respectively, from average and above-average growth.

10. Relative to the comparator countries, Italy’s average annual real GDP growth during 1990-98 lagged to varying degrees (Table 1). Lower working-age population growth (Table 2) may have played a role, but a per (working age) capita growth differential of 1 percentage point remains relative to the Netherlands, and of ⅓–¾ percentage point relative the United Kingdom depending on whether the focus is on total or business GDP, respectively. However, demographic factors are unlikely to have been a major constraint on growth, as labor force participation did not rise, nor did unemployment decline.

Table 1.

Italy: Growth Performance, 1980-99

Source: IMF, World Economic Outlook database, unless otherwise noted.

^1/Source: OECD Economic Outlook database.

Table 1.

Italy: Growth Performance, 1980-99

	1980-89	1990-99	1996-99
(Annual percentage change, unless otherwise indicated)
Real GDP
France	2.4	1.7	2.3
Italy	2.4	1.4	1.5
Netherlands	1.9	2.9	3.5
United Kingdom	2.4	1.9	2.6
Real GDP, business 1/
France	2.3	1.8	2.5
Italy	2.5	1.5	1.4
Netherlands	1.9	3.0	3.7
United Kingdom	2.9	2.5	2.8
Employment 1/
France	0.2	0.3	0.9
Italy	0.2	−0.1	0.8
Netherlands	0.7	2.1	2.7
United Kingdom	0.6	0.2	1.2
Government employment 1/
France	1.5	0.8	0.3
Italy	1.4	−0.3	−0.5
Netherlands	0.6	−0.1	0.3
United Kingdom	−0.2	−3.6	−0.4
Capital stock business sector 1/
France	2.7	2.4	2.2
Italy	2.9	2.8	2.7
Netherlands	1.7	2.4	2.8
United Kingdom	1.7	2.8	3.1

Source: IMF, World Economic Outlook database, unless otherwise noted.

^1/Source: OECD Economic Outlook database.

View Table

Table 1.

Italy: Growth Performance, 1980-99

	1980-89	1990-99	1996-99
(Annual percentage change, unless otherwise indicated)
Real GDP
France	2.4	1.7	2.3
Italy	2.4	1.4	1.5
Netherlands	1.9	2.9	3.5
United Kingdom	2.4	1.9	2.6
Real GDP, business 1/
France	2.3	1.8	2.5
Italy	2.5	1.5	1.4
Netherlands	1.9	3.0	3.7
United Kingdom	2.9	2.5	2.8
Employment 1/
France	0.2	0.3	0.9
Italy	0.2	−0.1	0.8
Netherlands	0.7	2.1	2.7
United Kingdom	0.6	0.2	1.2
Government employment 1/
France	1.5	0.8	0.3
Italy	1.4	−0.3	−0.5
Netherlands	0.6	−0.1	0.3
United Kingdom	−0.2	−3.6	−0.4
Capital stock business sector 1/
France	2.7	2.4	2.2
Italy	2.9	2.8	2.7
Netherlands	1.7	2.4	2.8
United Kingdom	1.7	2.8	3.1

Source: IMF, World Economic Outlook database, unless otherwise noted.

^1/Source: OECD Economic Outlook database.

Table 2.

Italy: Demographics and Growth, 1980-99

Sources: IMF; OECD Economic Outlook database.

Table 2.

Italy: Demographics and Growth, 1980-99

	1980-89	1990-99	1996-99
Working age population growth
France	0.9	0.4	0.4
Italy	0.7	0.1	−0.1
Netherlands	1.1	0.5	0.4
United Kingdom	0.5	0.3	0.3
Labor force participation rate
France	67.0	67.0	67.4
Italy	59.2	57.7	57.6
Netherlands	57.0	61.6	63.9
United Kingdom	74.7	75.6	75.4
Unemployment rate
France	9.0	11.2	11.9
Italy	8.4	10.7	11.8
Netherlands	8.0	5.8	4.9
United Kingdom	9.5	8.0	6.8
Real GDP per person of working age
France	1.4	1.4	1.9
Italy	1.7	1.3	1.5
Netherlands	0.8	2.4	3.1
United Kingdom	1.9	1.6	2.2

Sources: IMF; OECD Economic Outlook database.

View Table

Table 2.

Italy: Demographics and Growth, 1980-99

	1980-89	1990-99	1996-99
Working age population growth
France	0.9	0.4	0.4
Italy	0.7	0.1	−0.1
Netherlands	1.1	0.5	0.4
United Kingdom	0.5	0.3	0.3
Labor force participation rate
France	67.0	67.0	67.4
Italy	59.2	57.7	57.6
Netherlands	57.0	61.6	63.9
United Kingdom	74.7	75.6	75.4
Unemployment rate
France	9.0	11.2	11.9
Italy	8.4	10.7	11.8
Netherlands	8.0	5.8	4.9
United Kingdom	9.5	8.0	6.8
Real GDP per person of working age
France	1.4	1.4	1.9
Italy	1.7	1.3	1.5
Netherlands	0.8	2.4	3.1
United Kingdom	1.9	1.6	2.2

Sources: IMF; OECD Economic Outlook database.

11. This chapter argues that structural impediments—and not mainly macroeconomic policies—have precluded Italy from growing at a pace similar to that of the Netherlands or the United Kingdom over the past decade. The starting point for investigating structural growth impediments is a standard growth accounting exercise. The results from the exercise suggest that Italy has been lagging the Netherlands and the United Kingdom mainly with respect to employment creation. Studying the labor market reveals that Italy suffered more adverse structural labor demand developments, which may have been related to insufficient wage dispersion and, during the 1990s, to enterprise restructuring. However, the main difference relative to the comparators lies in labor supply developments: considering relative labor market conditions, there was less moderation of gross wages in Italy than elsewhere. This is likely to have been related to developments in labor taxes. While moderation of pre-tax wages was supported through tax and contribution reductions in the Netherlands over the past two decades, labor taxes increased sharply in Italy.

12. The paper is structured as follows: Section A scrutinizes demand-side explanations for Italy’s underperformance. Section B elaborates on various structural hypotheses to explain Italy’s lower growth; and Section C concludes and draws policy lessons.

A. Demand-Side Developments

13. Demand-side policies, notably a tighter fiscal stance and higher interest rates than in the comparator countries, have been held accountable for Italy’s lower growth over the past decade. This section shows that under similar macroeconomic policies, Italy’s GDP growth would have been comparable to that of France over the 1990s. However, relative to the Netherlands and the United Kingdom, annual growth differentials of 1-1½ percentage points remain. Considering that absent a policy tightening, Italy’s euro participation would not have been feasible and macroeconomic policies not sustainable—with implications for risk premia and growth that are not well captured in the simulation exercise—the simulation results probably present an upper bound on the impact of policy tightening. Thus, structural rather than demand-side factors probably explain Italy’s lower GDP growth relative to these two countries.

14. A recent European Commission (EC) study (European Commission (2000)) focuses on a demand-side analysis of economic developments. It argues that a tighter fiscal policy only partly explains Italy’s lower growth. Had the primary fiscal balance been left unchanged starting in 1996—implying a 2½ percent and 1 percent of GDP lower surplus in 1997 and 1998, respectively—the EC study estimates that annual real GDP growth in 1997-98 would have been about ½ percentage point higher, leaving it still much lower than in the comparators.

15. This section explains how Italy’s real GDP growth might have differed, had it been able to adopt the same fiscal policies as the comparators had over 1992-98 and benefited from the same interest and exchange rate developments. While this exercise may not assess Italy’s growth under a viable counterfactual scenario, it is nonetheless essential to understand Italy’s growth performance during the 1990s.

16. Italy experienced the largest withdrawal of fiscal stimulus over 1992-98, as measured by the change in the primary structural fiscal balance (Figure 1): over this period Italy had the lowest and highest growth in structural primary expenditure and structural revenue, respectively. Both the size and the length of the adjustment effort likely have affected growth. Concerning monetary policy, the differences across countries are subtler. While Italy clearly had the highest real interest rates during the 1990s, it also benefited from the largest decline in long- and short-term interest rates, albeit mostly in the course of 1996-98. Both interest rate levels and changes matter for investment; however, to the extent that the stock of capital had adjusted to a level consistent with interest rates prevailing in 1990-91, interest rate changes may be more relevant.²

View Full Size

Italy: Fiscal and Financial Indicators, 1990s

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: WEO Database; and Fund staff calculations.

• Download Figure
• Download figure as PowerPoint slide

17. The question that then arises is how real GDP growth in Italy would have fared, if the country had adopted the same fiscal policy that the comparators did and benefited from the same interest rate and exchange rate changes during the 1990s. A “same fiscal policy” is defined as a policy that would have imparted the same fiscal impulse—as measured by the estimated change in the structural primary fiscal balance—as in the comparator countries.³ The differences between the Italian impulse on the one hand and that in the comparator countries on the other are assumed to be reflected entirely in higher or lower public consumption, depending on whether the fiscal impulse in Italy was smaller or larger, respectively.

18. From Table 3 it is possible to see how the fiscal impulse (and thus public consumption) and changes in interest rates and in the exchange rate would have differed for Italy, had they evolved as in the comparator countries. The table also displays the elasticities of the real GDP level with respect to these variables of the pre-EMU Oxford Economic Forecasting (OEF) Model.⁴ These elasticities are used to compute—in Tables 3-4—how GDP growth in Italy would have differed under the comparators’ fiscal policies, and interest and exchange rate changes, during 1992-98. Under the policies and conditions of France and the Netherlands, Italy’s real GDP would have grown faster by about ⅓ percentage point than it actually did; under the United Kingdom’s economic scenario; it would actually have grown some ⅓ percentage point slower. Fiscal policy mainly explains the growth differentials; only relative to the United Kingdom was the estimated impact of alternative monetary and exchange rate paths larger than for alternative fiscal policies.

Table 3.

Italy: Simulating Identical Fiscal and Monetary Policy Stance, 1992-98

Sources: Fund staff estimates.

Table 3.

Italy: Simulating Identical Fiscal and Monetary Policy Stance, 1992-98

Model assumptions (deviations from baseline)
	T	T+1	T+2	T+3	T+4	T+8	T+6
Elasticity of GDP w.r.t. government consumption +1 percent of GDP	0.9	0.7	0.7	0.7	0.6	0.4	0.3
Elasticity of GDP w.r.t. interest rate +1 percent	−0.3	−0.7	−0.9	−1.1	−1.2	−1.2	−1.2
Elasticity of GDP w.r.t. exchange rate +1 percent	−0.2	−0.1	−0.1	−0.1	0.0	0.0	0.0
Model input: fiscal and monetary conditions	1992	1993	1994	1995	1996	1997	1998	Average 1992-98	Actual 1992-98
Fiscal impulse (based on structural primary fiscal balance in percent of potential GDP):
France	1.6	−0.4	−0.1	0.0	−1.7	−0.8	0.6	−0.1	…
Italy	−2.4	−1.6	0.7	−1.6	−0.7	−2.5	1.2	−1.0	…
Netherlands	0.4	−1.8	0.9	−0.2	−1.9	0.2	0.3	−0.3	…
United Kingdom	1.3	0.8	−0.6	−0.2	−1.1	−2.0	−1.7	−0.5	…
Annual interest rate changes (average of short-term and long-term rates):
France	−0.1	−2.0	−0.7	−0.3	−1.2	−0.4	−0.2	−0.7	…
Italy	1.0	−3.0	−1.2	1.8	−2.5	−2.3	−1.9	−1.1	…
Netherlands	−0.3	−1.1	−1.5	−0.4	−1.1	−0.2	−0.4	−0.7	…
United Kingdom	−1.5	−2.6	0.2	0.6	−0.5	0.0	−0.5	−0.6	…
Annual nominal effective exchange rate changes:
France	3.2	3.5	0.9	3.0	−0.1	−3.2	0.5	1.1	…
Italy	−3.1	−15.8	−4.4	−9.8	9.3	0.7	−0.4	−3.6	…
Netherlands	2.3	3.1	0.3	4.0	−1.9	−4.6	−0.1	0.4	…
United Kingdom	−3.8	−8.3	0.3	−4.9	1.7	16.5	3.4	0.4	…
Model output: additional growth of Italy relative to comparator
Same fiscal stance:
France	3.6	0.3	−1.0	1.5	−1.6	1.6	−0.9	0.5	0.4
Netherlands	2.6	−0.7	0.2	1.2	−1.6	2.6	−1.4	0.4	1.4
United Kingdom	3.4	1.4	−1.7	1.5	−1.0	0.2	−2.6	0.2	1.3
Same interest rate changes:
France	0.3	0.1	−0.3	0.5	0.3	−0.9	−1.2	−0.2	0.4
Netherlands	0.4	−0.1	−0.4	0.7	0.2	−0.9	−1.1	−0.2	1.4
United Kingdom	0.8	0.9	−0.1	0.2	−0.2	−1.5	−1.6	−0.2	1.3
Same nominal exchange rate changes:
France	−1.1	−3.0	0.8	−1.4	3.5	1.0	0.4	0.0	0.4
Netherlands	−1.0	−3.0	0.9	−1.7	3.9	1.0	0.4	0.1	1.4
United Kingdom	0.1	−1.4	−0.2	−0.4	2.0	−3.2	1.0	−0.3	1.3

Sources: Fund staff estimates.

View Table

Table 3.

Italy: Simulating Identical Fiscal and Monetary Policy Stance, 1992-98

Model assumptions (deviations from baseline)
	T	T+1	T+2	T+3	T+4	T+8	T+6
Elasticity of GDP w.r.t. government consumption +1 percent of GDP	0.9	0.7	0.7	0.7	0.6	0.4	0.3
Elasticity of GDP w.r.t. interest rate +1 percent	−0.3	−0.7	−0.9	−1.1	−1.2	−1.2	−1.2
Elasticity of GDP w.r.t. exchange rate +1 percent	−0.2	−0.1	−0.1	−0.1	0.0	0.0	0.0
Model input: fiscal and monetary conditions	1992	1993	1994	1995	1996	1997	1998	Average 1992-98	Actual 1992-98
Fiscal impulse (based on structural primary fiscal balance in percent of potential GDP):
France	1.6	−0.4	−0.1	0.0	−1.7	−0.8	0.6	−0.1	…
Italy	−2.4	−1.6	0.7	−1.6	−0.7	−2.5	1.2	−1.0	…
Netherlands	0.4	−1.8	0.9	−0.2	−1.9	0.2	0.3	−0.3	…
United Kingdom	1.3	0.8	−0.6	−0.2	−1.1	−2.0	−1.7	−0.5	…
Annual interest rate changes (average of short-term and long-term rates):
France	−0.1	−2.0	−0.7	−0.3	−1.2	−0.4	−0.2	−0.7	…
Italy	1.0	−3.0	−1.2	1.8	−2.5	−2.3	−1.9	−1.1	…
Netherlands	−0.3	−1.1	−1.5	−0.4	−1.1	−0.2	−0.4	−0.7	…
United Kingdom	−1.5	−2.6	0.2	0.6	−0.5	0.0	−0.5	−0.6	…
Annual nominal effective exchange rate changes:
France	3.2	3.5	0.9	3.0	−0.1	−3.2	0.5	1.1	…
Italy	−3.1	−15.8	−4.4	−9.8	9.3	0.7	−0.4	−3.6	…
Netherlands	2.3	3.1	0.3	4.0	−1.9	−4.6	−0.1	0.4	…
United Kingdom	−3.8	−8.3	0.3	−4.9	1.7	16.5	3.4	0.4	…
Model output: additional growth of Italy relative to comparator
Same fiscal stance:
France	3.6	0.3	−1.0	1.5	−1.6	1.6	−0.9	0.5	0.4
Netherlands	2.6	−0.7	0.2	1.2	−1.6	2.6	−1.4	0.4	1.4
United Kingdom	3.4	1.4	−1.7	1.5	−1.0	0.2	−2.6	0.2	1.3
Same interest rate changes:
France	0.3	0.1	−0.3	0.5	0.3	−0.9	−1.2	−0.2	0.4
Netherlands	0.4	−0.1	−0.4	0.7	0.2	−0.9	−1.1	−0.2	1.4
United Kingdom	0.8	0.9	−0.1	0.2	−0.2	−1.5	−1.6	−0.2	1.3
Same nominal exchange rate changes:
France	−1.1	−3.0	0.8	−1.4	3.5	1.0	0.4	0.0	0.4
Netherlands	−1.0	−3.0	0.9	−1.7	3.9	1.0	0.4	0.1	1.4
United Kingdom	0.1	−1.4	−0.2	−0.4	2.0	−3.2	1.0	−0.3	1.3

Sources: Fund staff estimates.

Table 4.

Italy: Decomposing Growth Differentials, 1992-98

Source: Fund staff estimates.

Table 4.

Italy: Decomposing Growth Differentials, 1992-98

			Average 1992-98 Model	Average 1996-98 Model	Average 1992-98 Actual	Average 1996-98 Actual
Additional growth of Italy relative to comparator countries:
France			0.4	0.8	0.2	0.7
	of which:	fiscal impulse	0.5	−0.3
		monetary policy	−0.2	−0.6
		exchange rate	0.0	1.6
Netherlands			0.3	1.0	1.3	2.0
	of which:	fiscal impulse	0.4	−0.1
		monetary policy	−0.2	−0.6
		exchange rate	0.1	1.8
United Kingdom			−0.4	−2.3	1.2	1.3
	of which:	fiscal impulse	0.2	−1.1
		monetary policy	−0.2	−1.1
		exchange rate	−0.3	−0.1

Source: Fund staff estimates.

View Table

Table 4.

Italy: Decomposing Growth Differentials, 1992-98

			Average 1992-98 Model	Average 1996-98 Model	Average 1992-98 Actual	Average 1996-98 Actual
Additional growth of Italy relative to comparator countries:
France			0.4	0.8	0.2	0.7
	of which:	fiscal impulse	0.5	−0.3
		monetary policy	−0.2	−0.6
		exchange rate	0.0	1.6
Netherlands			0.3	1.0	1.3	2.0
	of which:	fiscal impulse	0.4	−0.1
		monetary policy	−0.2	−0.6
		exchange rate	0.1	1.8
United Kingdom			−0.4	−2.3	1.2	1.3
	of which:	fiscal impulse	0.2	−1.1
		monetary policy	−0.2	−1.1
		exchange rate	−0.3	−0.1

Source: Fund staff estimates.

19. The results can also be used to evaluate the growth performance over 1996-98. The model suggests that under similar fiscal policies and monetary conditions, average annual growth in Italy would have been higher by ¾ and 1 percentage point, respectively, relative to France and the Netherlands; by contrast, growth would have been considerably lower relative to the United Kingdom. Relative to France and the Netherlands, the growth differentials during this period largely reflect the lira’s appreciation since 1996.

20. In evaluating the simulation results, it is important to bear in mind the limitations of this exercise. First, the OEF model, which was used in the exercise, may overestimate the short- run effects of changes in policies and monetary conditions on output.⁵ Second, if Italy had been able to meet the Maastricht fiscal deficit criterion with discretionary policy changes of the magnitude of such changes in France or the Netherlands, it might have chosen to forego much of the increase in fiscal revenues, which is where most of the fiscal adjustment during the 1990s was achieved, rather than allocate additional expenditure to public consumption. Third, and perhaps most important, pursuing policies comparable to those adopted by France, the Netherlands, and the United Kingdom was not an option for Italy without foregoing the benefits of euro entry: interest rate premia would then likely have been higher, with adverse effects on growth. Bearing these caveats in mind, even the simulations results suggest that structural impediments—and not mainly macroeconomic policies—precluded Italy from growing at a pace similar to that of the Netherlands or the United Kingdom over the past decade: most of Italy’s growth differential during 1992-98 relative to these countries cannot be accounted for by macroeconomic policies and developments under the OEF model.

B. Structural Impediments

21. Various structural features of the Italian economy have been considered responsible for the country’s underperformance. At the risk of oversimplifying, they can be grouped as follows: (i) low productivity growth, owing to distorted product markets (Table 5), as well as a lack of education (Table 6) and innovation (Table 7); (ii) high labor (Table 8) and capital income (Table 9) taxes, which hold back employment and capital accumulation; (iii) distortive labor market regulation (Table 10); and (iv) rigidities in the Mezzogiorno, the poorer south of Italy, which hold back growth at the national level (Table 11).

Table 5.

Italy: Product Market Regulation

Source: Nicoletti and others (1999). The ranking spans 0-6, with 6 indicating the most restrictive level. Rankings were determined on the basis of responses to an extensive questionnaire distributed to OECD member countries in 1998. The questionnaire asked for information on more than 1,500 regulatory provisions. The sectors included retail distribution, transportation (road freight, air passenger transport, and rail transport), and telecommunications.

Table 5.

Italy: Product Market Regulation

		France	Italy	Netherlands	United Kingdom
Overall
Product market regulation		2.1	2.3	1.4	0.5
Outward-oriented regulation		1.2	0.8	0.7	0.4
Inward-oriented regulation		2.7	3.3	1.8	0.5
A. Thematic
	State control	2.6	3.9	2.3	0.6
	Barriers to entrepreneurship	2.7	2.7	1.4	0.5
B. Functional
	Economic regulation	2.3	3.5	2.1	0.6
	Administrative regulation	3.1	3.0	1.5	0.5

Source: Nicoletti and others (1999). The ranking spans 0-6, with 6 indicating the most restrictive level. Rankings were determined on the basis of responses to an extensive questionnaire distributed to OECD member countries in 1998. The questionnaire asked for information on more than 1,500 regulatory provisions. The sectors included retail distribution, transportation (road freight, air passenger transport, and rail transport), and telecommunications.

View Table

Table 5.

Italy: Product Market Regulation

		France	Italy	Netherlands	United Kingdom
Overall
Product market regulation		2.1	2.3	1.4	0.5
Outward-oriented regulation		1.2	0.8	0.7	0.4
Inward-oriented regulation		2.7	3.3	1.8	0.5
A. Thematic
	State control	2.6	3.9	2.3	0.6
	Barriers to entrepreneurship	2.7	2.7	1.4	0.5
B. Functional
	Economic regulation	2.3	3.5	2.1	0.6
	Administrative regulation	3.1	3.0	1.5	0.5

Source: Nicoletti and others (1999). The ranking spans 0-6, with 6 indicating the most restrictive level. Rankings were determined on the basis of responses to an extensive questionnaire distributed to OECD member countries in 1998. The questionnaire asked for information on more than 1,500 regulatory provisions. The sectors included retail distribution, transportation (road freight, air passenger transport, and rail transport), and telecommunications.

Table 6.

Italy: Educational Attainment and Expenditure, 1995-96

Source: Modified versions of Tables A1.2a and B4.1 from OECD (1998).

^1/Converted In U.S. dollars using PPPs, on public and private institutions (based on full-time equivalents).

^2/Public institutions.

Table 6.

Italy: Educational Attainment and Expenditure, 1995-96

	At least upper secondary education, 1996					At least university-level education, 1996
	Ages 25-64	Ages 25-34	Ages 35-44	Ages 45-54	Ages 55-64	Ages 25-64	Ages 25-34	Ages 35-44	Ages 45-54	Ages 55-64
France	60	74	64	56	38	10	12	10	10	5
Italy	38	52	46	31	17	8	8	11	8	5
Netherlands	63	72	66	57	47	23	25	25	21	16
United Kingdom	76	87	81	71	60	13	15	15	12	8
	Expenditure per student, 1995 1/
	Child	Primary	Secondary	Tertiary	Combined
France	3,242	3,379	6,182	6,569	5,001
Italy 1/	3,316	4,673	5,348	5,013	5,157
Netherlands	3,021	3,191	4,351	9,026	4,397
United Kingdom 2/	5,049	3,328	4,246	7,225	4,222

Source: Modified versions of Tables A1.2a and B4.1 from OECD (1998).

^1/Converted In U.S. dollars using PPPs, on public and private institutions (based on full-time equivalents).

^2/Public institutions.

View Table

Table 6.

Italy: Educational Attainment and Expenditure, 1995-96

	At least upper secondary education, 1996					At least university-level education, 1996
	Ages 25-64	Ages 25-34	Ages 35-44	Ages 45-54	Ages 55-64	Ages 25-64	Ages 25-34	Ages 35-44	Ages 45-54	Ages 55-64
France	60	74	64	56	38	10	12	10	10	5
Italy	38	52	46	31	17	8	8	11	8	5
Netherlands	63	72	66	57	47	23	25	25	21	16
United Kingdom	76	87	81	71	60	13	15	15	12	8
	Expenditure per student, 1995 1/
	Child	Primary	Secondary	Tertiary	Combined
France	3,242	3,379	6,182	6,569	5,001
Italy 1/	3,316	4,673	5,348	5,013	5,157
Netherlands	3,021	3,191	4,351	9,026	4,397
United Kingdom 2/	5,049	3,328	4,246	7,225	4,222

Source: Modified versions of Tables A1.2a and B4.1 from OECD (1998).

^1/Converted In U.S. dollars using PPPs, on public and private institutions (based on full-time equivalents).

^2/Public institutions.

Table 7.

Italy: Selected R&D Indicators, 1997-98

Source: Eurostat.

Table 7.

Italy: Selected R&D Indicators, 1997-98

Budget appropriations (in percent of GDP), 1997
France	1.05
Italy	0.62
Netherlands	0.82
United Kingdom	0.75
Employment in high-tech sectors (in percent of total), 1998
France	10.6
Italy	10.1
Netherlands	8.2
United Kingdom	11.5
European patent applications (per million persons), average 1989-98
	Level	Growth Rate
France	95.5	1.9
Italy	45.9	4.1
Netherlands	121.1	2.3
United Kingdom	78.8	0.3

Source: Eurostat.

View Table

Table 7.

Italy: Selected R&D Indicators, 1997-98

Budget appropriations (in percent of GDP), 1997
France	1.05
Italy	0.62
Netherlands	0.82
United Kingdom	0.75
Employment in high-tech sectors (in percent of total), 1998
France	10.6
Italy	10.1
Netherlands	8.2
United Kingdom	11.5
European patent applications (per million persons), average 1989-98
	Level	Growth Rate
France	95.5	1.9
Italy	45.9	4.1
Netherlands	121.1	2.3
United Kingdom	78.8	0.3

Source: Eurostat.

Table 8.

Italy: Labor Taxation, 1980-96

Source: OECD (1999a and 1999b).

^1/These effective tax rates are based on measured tax revenue and income. Average effective tax rates on labor are estimated as household income taxes and social security contributions of employees, employers, and the self-employed divided by total compensation of employees plus self-employed labor income.

^2/In percent of labor costs, for a single person without children. For Italy, as of 1990, data have been revised to include only production workers (excluding employees). Also, the decline in the tax wedge in 1999 reflects to an important extent the replacement of employer-paid health contributions (and various other smaller taxes) by a tax on firms’ value added (IRAP).

^3/Estimates.

Table 8.

Italy: Labor Taxation, 1980-96

Average effective tax rates on labor, 1980-96 1/
	1980	1985	1990	1996
France	36	40	42	45
Italy	26	30	33	39
Netherlands	44	47	47	44
United Kingdom	25	24	22	22
Income tax plus employee and employee contribution 2/
	1979	1985	1991	1994	1996	1999 3/
France	…	…	…	51.6	49.7	47.9
Italy	45.3	50.0	48.8	49.9	50.8	47.3
Netherlands	48.0	49.9	46.5	45.6	43.8	44.4
United Kingdom	36.1	37.8	33.2	33.3	32.6	31.0
Marginal tax wedges on earned income-selected income levels, 1996
	Single-no children			Head earner-2 children			Spouse-2 children
Wage level in percent of average production worker	67	100	167	100-0	100-33	100-67	100-0	100-33	100-67
France	52.1	55.1	54.9	46.4	45.4	49.6	43.2	43.2	47.7
Italy	55.1	59.4	57.6	59.4	59.4	59.4	38.4	52.0	55.1
Netherlands	52.2	57.5	50.0	46.9	57.5	57.5	11.1	42.7	52.2
United Kingdom	38.3	40.1	31.0	40.1	40.1	40.1	0.0	32.0	38.3

Source: OECD (1999a and 1999b).

^1/These effective tax rates are based on measured tax revenue and income. Average effective tax rates on labor are estimated as household income taxes and social security contributions of employees, employers, and the self-employed divided by total compensation of employees plus self-employed labor income.

^2/In percent of labor costs, for a single person without children. For Italy, as of 1990, data have been revised to include only production workers (excluding employees). Also, the decline in the tax wedge in 1999 reflects to an important extent the replacement of employer-paid health contributions (and various other smaller taxes) by a tax on firms’ value added (IRAP).

^3/Estimates.

View Table

Table 8.

Italy: Labor Taxation, 1980-96

Average effective tax rates on labor, 1980-96 1/
	1980	1985	1990	1996
France	36	40	42	45
Italy	26	30	33	39
Netherlands	44	47	47	44
United Kingdom	25	24	22	22
Income tax plus employee and employee contribution 2/
	1979	1985	1991	1994	1996	1999 3/
France	…	…	…	51.6	49.7	47.9
Italy	45.3	50.0	48.8	49.9	50.8	47.3
Netherlands	48.0	49.9	46.5	45.6	43.8	44.4
United Kingdom	36.1	37.8	33.2	33.3	32.6	31.0
Marginal tax wedges on earned income-selected income levels, 1996
	Single-no children			Head earner-2 children			Spouse-2 children
Wage level in percent of average production worker	67	100	167	100-0	100-33	100-67	100-0	100-33	100-67
France	52.1	55.1	54.9	46.4	45.4	49.6	43.2	43.2	47.7
Italy	55.1	59.4	57.6	59.4	59.4	59.4	38.4	52.0	55.1
Netherlands	52.2	57.5	50.0	46.9	57.5	57.5	11.1	42.7	52.2
United Kingdom	38.3	40.1	31.0	40.1	40.1	40.1	0.0	32.0	38.3

Source: OECD (1999a and 1999b).

^1/These effective tax rates are based on measured tax revenue and income. Average effective tax rates on labor are estimated as household income taxes and social security contributions of employees, employers, and the self-employed divided by total compensation of employees plus self-employed labor income.

^2/In percent of labor costs, for a single person without children. For Italy, as of 1990, data have been revised to include only production workers (excluding employees). Also, the decline in the tax wedge in 1999 reflects to an important extent the replacement of employer-paid health contributions (and various other smaller taxes) by a tax on firms’ value added (IRAP).

^3/Estimates.

Table 9.

Italy: Capital Taxation, 1980-98

Sources: OECD (2000a); and Carey and others (2000), forthcoming OECD Working Paper.

^1/The average effective tax rates are based on measured tax revenue and income. Average effective tax rates on capital are estimated as household taxes (including unallocated social security contributions between employees, employers, and the self-employed) paid on self-employment and property income received by households less imputed wages for the self-employed, plus corporate taxes, taxes on property and divided by the operating surplus of the economy less wages of the self-employed.

^2/These indicators show the degree to which the personal and corporate tax systems scale up (or down) the real pre-tax rate of return that must be earned on an investment, given that the household can earn a 5 percent real rate of return on a demand deposit.

^3/The weights are: machinery 0.50, buildings 0.28, and inventories 0.22.

^4/The weights are: retained earnings 0.55, new equity 0.10, and debt 0.35.

Table 9.

Italy: Capital Taxation, 1980-98

Average effective tax rates on capital (based on gross operating surplus), 1980-96 1/
	1980	1985	1990	1996
France	21	22	21	20
Italy	19	22	26	30
Netherlands	21	16	19	21
United Kingdom	43	46	42	39
Marginal tax wedges on investment, personal, and corporate systems, 1991-98 2/
Country	Year	Sources of financing 3/			Physical assets 4/			Overall weighted average
		Earnings	Equity	Debt	Machinery	Building	Inventories
France	1991	2.7	6.4	−0.1	1.2	2.1	4.0	2.1
	1996	3.5	6.2	…	2.3	2.4	2.1	2.3
	1998	4.4	8.5	0.8	2.6	4.1	4.8	3.5
Italy	1991	4.5	2.9	−0.8	2.6	3.1	1.7	2.5
	1996	5.9	5.3	…	2.6	3.4	6.9	3.8
	1998	2.2	2.5	0.6	1.0	1.8	3.1	1.7
Netherlands	1991	0.5	6.5	2.9	1.8	2.2	2.0	1.9
	1996	0.5	6.2	…	1.7	2.2	2.0	1.9
	1998	0.6	6.3	3.0	1.8	2.3	2.0	2.0
United Kingdom	1991	1.9	1.3	2.4	1.5	1.6	3.8	2.0
	1996	2.3	3.1	…	1.8	2.1	3.4	2.2
	1998	2.2	2.8	1.8	1.7	2.1	3.1	2.2

Sources: OECD (2000a); and Carey and others (2000), forthcoming OECD Working Paper.

^1/The average effective tax rates are based on measured tax revenue and income. Average effective tax rates on capital are estimated as household taxes (including unallocated social security contributions between employees, employers, and the self-employed) paid on self-employment and property income received by households less imputed wages for the self-employed, plus corporate taxes, taxes on property and divided by the operating surplus of the economy less wages of the self-employed.

^2/These indicators show the degree to which the personal and corporate tax systems scale up (or down) the real pre-tax rate of return that must be earned on an investment, given that the household can earn a 5 percent real rate of return on a demand deposit.

^3/The weights are: machinery 0.50, buildings 0.28, and inventories 0.22.

^4/The weights are: retained earnings 0.55, new equity 0.10, and debt 0.35.

View Table

Table 9.

Italy: Capital Taxation, 1980-98

Average effective tax rates on capital (based on gross operating surplus), 1980-96 1/
	1980	1985	1990	1996
France	21	22	21	20
Italy	19	22	26	30
Netherlands	21	16	19	21
United Kingdom	43	46	42	39
Marginal tax wedges on investment, personal, and corporate systems, 1991-98 2/
Country	Year	Sources of financing 3/			Physical assets 4/			Overall weighted average
		Earnings	Equity	Debt	Machinery	Building	Inventories
France	1991	2.7	6.4	−0.1	1.2	2.1	4.0	2.1
	1996	3.5	6.2	…	2.3	2.4	2.1	2.3
	1998	4.4	8.5	0.8	2.6	4.1	4.8	3.5
Italy	1991	4.5	2.9	−0.8	2.6	3.1	1.7	2.5
	1996	5.9	5.3	…	2.6	3.4	6.9	3.8
	1998	2.2	2.5	0.6	1.0	1.8	3.1	1.7
Netherlands	1991	0.5	6.5	2.9	1.8	2.2	2.0	1.9
	1996	0.5	6.2	…	1.7	2.2	2.0	1.9
	1998	0.6	6.3	3.0	1.8	2.3	2.0	2.0
United Kingdom	1991	1.9	1.3	2.4	1.5	1.6	3.8	2.0
	1996	2.3	3.1	…	1.8	2.1	3.4	2.2
	1998	2.2	2.8	1.8	1.7	2.1	3.1	2.2

Sources: OECD (2000a); and Carey and others (2000), forthcoming OECD Working Paper.

^1/The average effective tax rates are based on measured tax revenue and income. Average effective tax rates on capital are estimated as household taxes (including unallocated social security contributions between employees, employers, and the self-employed) paid on self-employment and property income received by households less imputed wages for the self-employed, plus corporate taxes, taxes on property and divided by the operating surplus of the economy less wages of the self-employed.

^2/These indicators show the degree to which the personal and corporate tax systems scale up (or down) the real pre-tax rate of return that must be earned on an investment, given that the household can earn a 5 percent real rate of return on a demand deposit.

^3/The weights are: machinery 0.50, buildings 0.28, and inventories 0.22.

^4/The weights are: retained earnings 0.55, new equity 0.10, and debt 0.35.

Table 10a.

Italy: Labor Market Regulation

Source: OECD (1999c). The summary scores can range from 0 to 6, with higher values representing more strict regulation.

^1/For late 1990s but excluding collective dismissal regulation.

^2/For late 1990s including collective dismissal regulation.

Table 10a.

Italy: Labor Market Regulation

	Late 1980s	Late 1990s
Strictness of protection against individual dismissals
France	2.5	2.6
Italy	3.1	3.1
Netherlands	3.2	3.4
United Kingdom	1.2	1.2
Strictness of temporary and fixed-term employment regulation
France	2.6	3.6
Italy	5.4	3.8
Netherlands	2.3	1.4
United Kingdom	0.3	0.3
Increase in strictness for collective relative to individual dismissals
France	…	2.1
Italy	…	4.1
Netherlands	…	2.8
United Kingdom	…	2.9
Overall strictness of employment protection legislation
		1/	2/
France	2.6	3.1	2.9
Italy	4.2	3.4	3.5
Netherlands	2.7	2.4	2.5
United Kingdom	0.7	0.7	1.1

Source: OECD (1999c). The summary scores can range from 0 to 6, with higher values representing more strict regulation.

^1/For late 1990s but excluding collective dismissal regulation.

^2/For late 1990s including collective dismissal regulation.

View Table

Table 10a.

Italy: Labor Market Regulation

	Late 1980s	Late 1990s
Strictness of protection against individual dismissals
France	2.5	2.6
Italy	3.1	3.1
Netherlands	3.2	3.4
United Kingdom	1.2	1.2
Strictness of temporary and fixed-term employment regulation
France	2.6	3.6
Italy	5.4	3.8
Netherlands	2.3	1.4
United Kingdom	0.3	0.3
Increase in strictness for collective relative to individual dismissals
France	…	2.1
Italy	…	4.1
Netherlands	…	2.8
United Kingdom	…	2.9
Overall strictness of employment protection legislation
		1/	2/
France	2.6	3.1	2.9
Italy	4.2	3.4	3.5
Netherlands	2.7	2.4	2.5
United Kingdom	0.7	0.7	1.1

Source: OECD (1999c). The summary scores can range from 0 to 6, with higher values representing more strict regulation.

^1/For late 1990s but excluding collective dismissal regulation.

^2/For late 1990s including collective dismissal regulation.

Table 10b.

Italy: Labor Market Regulation, 1989-94

Source: Nickell (1997)

Table 10b.

Italy: Labor Market Regulation, 1989-94

	1989-94
Benefit replacement rate (in percent)
France	57.0
Italy	20.0
Netherlands	70.0
United Kingdom	38.0
Benefit duration (in years)
France	3.0
Italy	0.5
Netherlands	2.0
United Kingdom	4.0

Source: Nickell (1997)

View Table

Table 10b.

Italy: Labor Market Regulation, 1989-94

	1989-94
Benefit replacement rate (in percent)
France	57.0
Italy	20.0
Netherlands	70.0
United Kingdom	38.0
Benefit duration (in years)
France	3.0
Italy	0.5
Netherlands	2.0
United Kingdom	4.0

Source: Nickell (1997)

Table 11.

Italy: Center-North and South, 1980-99

Sources: SVIMEZ for 1996-98; and ISTAT otherwise. Data are based on ESA79.

Table 11.

Italy: Center-North and South, 1980-99

	1981-89	1990-99	1996-99
(Annual percentage change, unless otherwise indicated)
Real GDP
Italy	2.2	1.3	1.2
Center-North	2.2	1.4	1.4
South	2.3	0.8	0.7
Real GDP, business
Italy	2.5	1.4	1.2
Center-North	2.5	1.6	1.4
South	2.5	0.9	0.6
Employment
Italy	0.5	−0.4	0.2
Center-North	0.5	−0.3	0.2
South	0.7	−0.7	0.0
Employment, nongovernment
Italy	0.3	−0.5	0.3
Center-North	0.3	−0.3	0.3
South	0.3	−0.9	0.2
Gross fixed capital formation (in percent of GDP)
Italy	19.6	18.5	17.8
Center-North	18.6	18.4	18.2
South	22.7	18.5	16.5
	(In percent of Center- North)
	1980	1990	1998
Nominal Per capita GDP of South	57.8	57.5	54.6
Real GDP of South	33.2	33.3	31.5
Population of South	54.8	57.9	58.4

Sources: SVIMEZ for 1996-98; and ISTAT otherwise. Data are based on ESA79.

View Table

Table 11.

Italy: Center-North and South, 1980-99

	1981-89	1990-99	1996-99
(Annual percentage change, unless otherwise indicated)
Real GDP
Italy	2.2	1.3	1.2
Center-North	2.2	1.4	1.4
South	2.3	0.8	0.7
Real GDP, business
Italy	2.5	1.4	1.2
Center-North	2.5	1.6	1.4
South	2.5	0.9	0.6
Employment
Italy	0.5	−0.4	0.2
Center-North	0.5	−0.3	0.2
South	0.7	−0.7	0.0
Employment, nongovernment
Italy	0.3	−0.5	0.3
Center-North	0.3	−0.3	0.3
South	0.3	−0.9	0.2
Gross fixed capital formation (in percent of GDP)
Italy	19.6	18.5	17.8
Center-North	18.6	18.4	18.2
South	22.7	18.5	16.5
	(In percent of Center- North)
	1980	1990	1998
Nominal Per capita GDP of South	57.8	57.5	54.6
Real GDP of South	33.2	33.3	31.5
Population of South	54.8	57.9	58.4

Sources: SVIMEZ for 1996-98; and ISTAT otherwise. Data are based on ESA79.

22. Growth accounting may help discriminate between various explanatory factors. Decomposing business sector real GDP growth reveals that Italy has lagged largely with respect to employment creation (Table 12). Problems in product markets, education, and research and the production structure would probably have been reflected in total factor productivity (TFP). TFP growth was somewhat smaller than in the comparator countries, except for the Netherlands; but the slowdown in productivity growth during the 1990s is difficult to rationalize. The product and services market environment clearly improved during this period, possibly even more rapidly in Italy than in most of the comparator countries. However, before the reforms began, Italy was lagging considerably behind the United Kingdom and the Netherlands. Similarly, no compelling case can be made that Italy has been falling behind with respect to education and research and development: on the contrary, the evidence suggests that Italy is catching up.

Table 12.

Italy: Growth Accounting, 1980-98 1/

Source: OECD; for 1997 no data are available for the United Kingdom. GDP stands for real GDP, TFP for factor productivity, L for employment, and K for capital, all in the business sector.

Table 12.

Italy: Growth Accounting, 1980-98 1/

	GDP	TFP	L	K
1990-98
France	1.7	0.8	0.0	0.9
Italy	1.5	0.7	−0.2	1.0
Netherlands	3.0	1.0	1.1	0.9
United Kingdom	2.6	1.0	0.7	0.9
1980-89
France	2.3	1.7	−0.1	0.8
Italy	2.5	1.2	0.4	0.9
Netherlands	1.9	1.0	0.2	0.6
United Kingdom	2.9	1.9	0.4	0.6

Source: OECD; for 1997 no data are available for the United Kingdom. GDP stands for real GDP, TFP for factor productivity, L for employment, and K for capital, all in the business sector.

View Table

Table 12.

Italy: Growth Accounting, 1980-98 1/

	GDP	TFP	L	K
1990-98
France	1.7	0.8	0.0	0.9
Italy	1.5	0.7	−0.2	1.0
Netherlands	3.0	1.0	1.1	0.9
United Kingdom	2.6	1.0	0.7	0.9
1980-89
France	2.3	1.7	−0.1	0.8
Italy	2.5	1.2	0.4	0.9
Netherlands	1.9	1.0	0.2	0.6
United Kingdom	2.9	1.9	0.4	0.6

Source: OECD; for 1997 no data are available for the United Kingdom. GDP stands for real GDP, TFP for factor productivity, L for employment, and K for capital, all in the business sector.

23. Turning to capital accumulation, the results show neither major differences between Italy and the comparators, nor a change in the contribution of capital over time. Several comments are in order: First, much of the slowdown in investment in Italy was in the public sector, and in construction, particularly for residential purposes, as indicated by a comparison of private nonresidential fixed investment with total gross fixed capital formation (Table 13). Second, while Italy stands out in having recorded a large increase in capital income taxes during 1980-96, the burden on capital was not necessarily higher than in the comparator countries over most of the period, and tax rates have recently come down considerably. Third, the results of the growth accounting exercise may be tainted by measurement error. Using other measures for factor inputs reveals that the contribution of capital to growth was declining in Italy during the 1990s (Box 1), although not nearly to an extent that would invalidate the fundamental result of the accounting exercise. Italy’s low output growth is closely related to developments in employment, a hypothesis that finds support in the evolution of the capital-output ratios (Figure 2).

Table 13.

Italy: Investment, 1980-99

Source: OECD.

Table 13.

Italy: Investment, 1980-99

	1980-89	1990-99	1996-99
Capital stock in the business sector (annual average growth)
France	2.7	2.4	2.2
Italy	2.9	2.8	2.7
Netherlands	1.7	2.4	2.8
United Kingdom	1.7	2.8	3.1
Private nonresidential fixed investment (in percent of business GDP)
France	15.2	15.7	15.3
Italy	14.4	14.9	15.0
Netherlands	15.9	16.6	17.1
United Kingdom	13.1	15.0	16.2
Private nonresidential fixed investment (in percent of capital)
France	5.7	5.7	5.5
Italy	5.3	5.1	5.0
Netherlands	5.8	6.6	6.9
United Kingdom	4.2	5.1	5.6
Private nonresidential fixed investment (contribution to business GDP growth)
France	0.5	0.2	0.6
Italy	0.3	0.2	0.4
Netherlands	0.4	0.5	1.0
United Kingdom	0.8	0.6	1.5
Gross fixed capital formation (contribution to GDP growth)
France	0.8	0.3	0.9
Italy	0.6	0.3	0.8
Netherlands	0.5	0.8	1.6
United Kingdom	0.9	0.5	1.5

Source: OECD.

View Table

Table 13.

Italy: Investment, 1980-99

	1980-89	1990-99	1996-99
Capital stock in the business sector (annual average growth)
France	2.7	2.4	2.2
Italy	2.9	2.8	2.7
Netherlands	1.7	2.4	2.8
United Kingdom	1.7	2.8	3.1
Private nonresidential fixed investment (in percent of business GDP)
France	15.2	15.7	15.3
Italy	14.4	14.9	15.0
Netherlands	15.9	16.6	17.1
United Kingdom	13.1	15.0	16.2
Private nonresidential fixed investment (in percent of capital)
France	5.7	5.7	5.5
Italy	5.3	5.1	5.0
Netherlands	5.8	6.6	6.9
United Kingdom	4.2	5.1	5.6
Private nonresidential fixed investment (contribution to business GDP growth)
France	0.5	0.2	0.6
Italy	0.3	0.2	0.4
Netherlands	0.4	0.5	1.0
United Kingdom	0.8	0.6	1.5
Gross fixed capital formation (contribution to GDP growth)
France	0.8	0.3	0.9
Italy	0.6	0.3	0.8
Netherlands	0.5	0.8	1.6
United Kingdom	0.9	0.5	1.5

Source: OECD.

View Full Size

Italy: Capital-Output Ratios, 1980-98

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: OECD; and Fund staff calculations.

• Download Figure
• Download figure as PowerPoint slide

24. Overall, growth accounting suggests that Italy’s subdued economic performance may be rooted in the labor market. This hypothesis appears as valid for the center-north of Italy as for the poorer South (Table 11) and is investigated further below. Note, though, that distorted product markets, less education and research, and lower capital accumulation may have directly affected the contribution of labor rather than TFP: growth accounting cannot offer firm conclusions on causal relationships.

Box 1.

Selected Data Issues

This box first proceeds to a growth accounting exercise for the business sector in Italy; however, unlike in the text, it excludes agriculture and energy. Second, it investigates the issue of the measurement of employment in terms of the number of employed (henceforth referred to as “bodies”) versus effective units.

Table A displays the results from growth accounting, obtained using the same shares of labor in business sector income as in the main text. The data for this growth accounting exercise differ in two important respects from the OECD data used in the text: (i) they do not include agriculture and energy; and (ii) the employment data are in effective units, basically the equivalent of man years. The results from the growth accounting exercise in the text are displayed as well. Compared with the exercise reported in the main text, the differences are most pronounced for the roles of employment and capital. The data for the entire business sector suggest a larger and smaller contribution of capital and labor, respectively. Also, for the 1990s they suggest a small increase rather than a decline in the contribution of capital to growth.

Table A.

Italy: Growth Accounting, 1980-98 ^1/

Source: Bank of Italy, unless otherwise noted.

^1/Data are for business sector, excluding agriculture and energy.

^2/Data are from OECD and cover the entire business sector.

Table A.

Italy: Growth Accounting, 1980-98 ^1/

	GDP	TFP	L	K
1990-98	1.6	0.9	0.1	0.7
1990-98 2/	1.5	0.7	−0.2	1.0
1980-89	2.7	1.2	0.6	0.9
1980-89 2/	2.5	1.2	0.4	0.9
1980-98	2.2	1.0	0.4	0.8
1980-98 2/	2.0	1.0	0.1	1.0

Source: Bank of Italy, unless otherwise noted.

^1/Data are for business sector, excluding agriculture and energy.

^2/Data are from OECD and cover the entire business sector.

View Table

Table A.

Italy: Growth Accounting, 1980-98 ^1/

	GDP	TFP	L	K
1990-98	1.6	0.9	0.1	0.7
1990-98 2/	1.5	0.7	−0.2	1.0
1980-89	2.7	1.2	0.6	0.9
1980-89 2/	2.5	1.2	0.4	0.9
1980-98	2.2	1.0	0.4	0.8
1980-98 2/	2.0	1.0	0.1	1.0

Source: Bank of Italy, unless otherwise noted.

^1/Data are for business sector, excluding agriculture and energy.

^2/Data are from OECD and cover the entire business sector.

Table B investigates the differences between the evolution of employment in effective units versus “bodies.” These data series can differ, for example, because of changing part-time employment or working time in general. In principle, it is preferable to use data in effective units rather than bodies in much of the work presented in this chapter. However, adjusting data in bodies to obtain effective units is hard because reliable data, particularly on actual working time, are difficult to obtain, even for the manufacturing sector. As can be seen, for Italy the average annual growth rate of business sector employment in units (ISTAT) is marginally lower than of employment in “bodies” (OECD). The opposite holds for the Netherlands, although the differences between data in units for the entire economy and data in bodies for the business sector are fairly small.

Table B.

Italy: Bodies and Units, 1981-95

Sources: For Italy, ISTAT; and SVIMEZ; for Netherlands, OECD; and Fund staff calculations.

^1/Refers to entire economy; all other data are for the business sector.

Table B.

Italy: Bodies and Units, 1981-95

1981-95
(Annual percentage change)
Italy
Bodies	−0.1
Units	−0.2
Netherlands
Units 1/	0.6
Bodies	0.7

Sources: For Italy, ISTAT; and SVIMEZ; for Netherlands, OECD; and Fund staff calculations.

^1/Refers to entire economy; all other data are for the business sector.

View Table

Table B.

Italy: Bodies and Units, 1981-95

1981-95
(Annual percentage change)
Italy
Bodies	−0.1
Units	−0.2
Netherlands
Units 1/	0.6
Bodies	0.7

Sources: For Italy, ISTAT; and SVIMEZ; for Netherlands, OECD; and Fund staff calculations.

^1/Refers to entire economy; all other data are for the business sector.

Data in effective units rather than “bodies” are preferable for computing labor costs. Labor costs are typically computed by dividing the wage bill through employment and, to the extent the evolution over time is interesting, dividing through some index of labor-augmenting technological progress (for example, total factor productivity scaled by the labor share). If employment is measured in bodies rather than effective units and bodies have grown much more rapidly than units (say because of a sharp increase in part-time employment, as experienced by the Netherlands), both the per capita labor cost and labor-augmenting technological progress could be underestimated. However, the effect of the former on productivity adjusted labor costs may dominate that of the latter.

Regardless of the data used in the growth accounting exercise, the key conclusion for Italy remains that the contribution of labor to growth was very low. Also, the data for employment in bodies and units for Italy and the Netherlands do not suggest that major biases result from using employment in bodies rather than units to compute productivity adjusted labor cost figures. Nevertheless, the role of working time is an issue that deserves further investigation (for a discussion of working time in manufacturing, see Casadio and D’Aurizio, 1999).

The labor market

25. Italy’s low growth appears closely linked to insufficient employment creation. Could the slow employment growth be related to labor market distortions and high labor costs? Answering this question requires an analysis of demand- and supply-side developments on the labor market.

26. Labor costs appear to have risen faster in Italy than in the comparator countries. Data for manufacturing labor costs (Bureau of Labor Statistics, 1999 and 2000) suggest that Italy experienced the largest cost increases since 1980 (Figure 3a). Similarly, real product wages in the broader business sector (Figure 3b)—these data underlie all subsequent analysis—grew considerably faster in Italy than in France and the Netherlands, albeit not more rapidly than in the United Kingdom.

View Full Size

Italy: Labor Costs, 1980-98

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: Bureau of Labor Statistics.

• Download Figure
• Download figure as PowerPoint slide

View Full Size

Italy: Real Wages, 1980-98 ^1/

(1990=100)

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: OECD; and Fund staff calculations.^1/ Computed using business sector GDP deflator.

• Download Figure
• Download figure as PowerPoint slide

27. However, labor costs alone tell little about potential distortions. A country with high real wage growth can have a less distorted labor market than a country with low growth. This can be the case if wages are driven by rising labor demand related, for example, to favorable productivity. To reach a more definite assessment of the labor market, several measures for determining structural labor supply and demand will be reviewed, following Blanchard (1998).

Labor supply

28. Labor supply is assumed to be characterized by a wage-setting function that relates real wages to unemployment rates.⁶ Assuming labor-augmenting technological progress,⁷ real wages w can grow at the rate of a, which is given by TFP scaled by the labor share in income α. Define the real effective wage as ω, with ω = log (w/a), then the wage relation is given by ω =-βU+Z, where U stands for the unemployment rate. Changes in Z can be thought of as shifts in the Blanchflower and Oswald (1994) “wage curve” and thus labor supply L^s. An increase in Z implies that workers ask for a higher real wage although unemployment has not declined and is equivalent to a drop in labor supply. The relation assumes reservation (and thus actual) wages rise with productivity: history clearly suggests no relation between productivity and employment levels or unemployment rates in the long run. However, if workers’ aspirations lag reality, reservation wages are unlikely to adjust instantaneously with TFP, giving rise to labor supply shocks in this model.

29. Is it sensible to build on such a wage relation? One perspective is offered by the job matching approach and wage bargaining.⁸ Workers cannot costlessly relocate to find a new job nor can they costlessly and instantaneously be replaced by their employers. In a depressed labor market, workers will settle for a wage close to the reservation wage because it is hard to find a job elsewhere. The opposite holds for a booming labor market. Alternatively, models focus on the firm-worker relationship, arguing that wages affect productivity. Firms may want to pay more to workers than their reservation wage to economize on turnover costs, motivate greater effort, or for social considerations, such as fairness.⁹

30. The following equation can thus be taken to characterize the evolution of labor supply L^S over time:

$\begin{array}{cc}{\mathrm{\Delta }L}_t^s={-\mathrm{\Delta }Z}_t=-\mathrm{\Delta }[{\omega }_t+{\beta U}_t]& \left(1\right)\end{array}$

With data for U, w, and α available, while a can be obtained from the growth accounting exercise (see Data Appendix for details), assessing labor supply developments requires an estimate for the parameter β. Blanchflower and Oswald (1995) investigate the relation between real wages and unemployment for three out of the four countries considered here (Italy, the Netherlands, and the United Kingdom). Using broadly comparable microeconomic data, they estimate a cross-sectional earnings equation for each country, in which, together with the usual set of control variables,¹⁰ the regional unemployment rate is entered as an explanatory variable. They demonstrate that there appears to be an empirical regularity in international pay and unemployment data, whereby estimates of the unemployment elasticity of pay cluster around -0.1. This result is broadly consistent with β = 1, considering the average unemployment rates in the countries, implying an increase in unemployment by 1 percentage point decreases effective wages by 1 percent. Note, however, that for Italy Brunello and others (1999), using a different model set-up and macroeconomic data, find evidence suggesting that β is closer to 2 rather than 1 in the long run.¹¹

31. The data suggest that for given labor market conditions, labor costs declined in all countries. Figure 4 plots a five-year moving average of Z to better capture structural changes: it can be thought of as the path of real effective wages at given unemployment rates in the various countries. The relation suggests that labor costs declined least in Italy, where effective wage growth did not slow during the second half of the 1980s, considering developments in unemployment. Relative to the Netherlands, for example, costs in Italy were scaled back some 15 percentage points less. Setting β = 2 changes this result little, although it now appears that labor costs have declined considerably less in France than under β = 1.

View Full Size

Italy: Labor Supply Developments, 1980-98 ^1/

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: OECD; and Fund staff calculations.^1/ Productivity-adjusted real wage at a constant unemployment or labor force participation rate.

• Download Figure
• Download figure as PowerPoint slide

32. A question that arises is whether the unemployment rate is the right measure of labor market conditions.¹² To the extent that the distinctions between being unemployed or out of the labor force are sometimes blurred—for example, owing to the discouraged worker effect or because of early retirement and disability support, measures that have been relied on in France, the Netherlands, or Italy to various degrees—the labor force participation rate (L/P) may better gauge labor market conditions. Nonetheless, substituting (1-L/P) for U in calculating the evolution of labor supply does not change the results noticeably for Italy although it does so for France.

Labor demand

33. How can labor demand developments be measured? Assume the following CES-type production:

$\begin{array}{cc}y=A{\left[\alpha {\left(an\right)}^{\frac{\sigma -1}{\sigma }}+\left(1-\alpha \right)k^{\frac{\sigma -1}{\sigma }}\right]}^{\frac{\sigma -1}{\sigma }},& \left(2\right)\end{array}$

where n stands for labor, k for capital, and σ for the (constant) elasticity of substitution between effective labor (an) and capital. To find labor demand, set the marginal product of labor equal to the real effective wage (w/a) multiplied by a markup (1+μ):

$\begin{array}{cc}\frac{\partial y}{\partial n}=\left(1+\mu \right)w& \left(3\right)\end{array}$

with the markup resulting, for example, from “efficient bargaining.”¹³ Take natural logarithms and rearrange to find:

$\begin{array}{cc}\log \left(1+\mu \right)=\mathrm{c}\mathrm{o}\mathrm{n}\mathrm{s}\tan \mathrm{t}+\log \alpha -\log \left(w/a\right)-\frac{1}{\sigma }\log \left(\frac{an}{y}\right)& \left(4\right)\end{array}$

34. There are two perspectives to changes in labor demand. A technology-driven decline in labor demand can be thought of as a drop in a for unchanged μ. Alternatively, a drop in demand because of less union bargaining power can be thought of as an increase in μ for given α. Observationally these changes in labor demand L^d are equivalent and their change over time is given by:

$\begin{array}{cc}\mathrm{\Delta }{L}_t^d=\mathrm{\Delta }\left[\log \left(w_t/a_t\right)+\frac{1}{\sigma }\log \left(\frac{a_t{n}_t}{y_t}\right)\right]& \left(5\right)\end{array}$

Conveniently, for σ = 1, the evolution of labor demand is given simply by the natural logarithm of the labor share in income (Figure 5).

View Full Size

Italy: Labor Demand Developments, 1980-98

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: OECD; and Fund staff calculations.

• Download Figure
• Download figure as PowerPoint slide

35. Note that equations (4) and (5) assume no costs of adjusting factor proportions. If it is costly for firms to adjust factor proportions, an increase in the wage will be associated with little contemporaneous change in (an/k) and thus (an/y); this in turn will lead to a decrease in the measured wedge. To allow for adjustment costs in factor proportions Blanchard (1998) proposes to replace log(w/a) with log(w_t^*) = 0.8 log w_t-1^* + 0.2 log (w_t/a_t) in computing L^d, implying a mean lag in the adjustment of factor proportions of four years:¹⁴ the results are fairly similar.

36. The evidence suggests that labor demand has declined in all countries except the United Kingdom over the past two decades (Figure 5). Upon allowing for costs in adjusting factor proportions and an elasticity of substitution less than 1, it is also clear that France and Italy suffered an about 5-10 percentage point larger demand decline—which was concentrated in the 1990s—than the Netherlands.

The regional dimension

37. Italy appears to have experienced the smallest increase in labor supply—meaning the smallest decline in labor costs for given labor market conditions—and, together with France, the largest decline in labor demand. A question that arises is whether developments in Italy were largely driven by the poor performance of the southern labor market. Evidence in Table 11 suggests that this might not have been the case: the weight of the South in the overall economy—equivalent to about one-fourth—is too small and cross-regional differences, while sizeable, insufficiently large to account for the central role in Italy’s relatively weak growth performance.

38. To explore the regional dimension of labor market development further, Figure 6 shows the evolution of labor demand and supply in the Center-North relative to that in the country as a whole. This exercise draws on a different data set and required the construction of regional capital stock data (see Data Appendix). For the latter reason, the results must be treated as only indicative. They suggest that labor demand evolved similarly across the country, a result that holds also for costs in adjusting factor proportions and σ = ¾. For labor supply, though, it appears that the reduction in labor costs in the Center-North was some 3-5 percentage points larger than in the country as a whole, depending on the value for β. The evolution of wage costs (and hence labor supply) of the Center-North was thus more like that of France.

View Full Size

Italy: Regional Labor Demand and Supply Differences, 1980-98

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: ISTAT; and Fund staff calculations.1/ Productivity-adjusted real wage at a constant unemployment rate.

• Download Figure
• Download figure as PowerPoint slide

39. In all, for given unemployment rates, labor costs declined considerably more in the Center-North than in the country as a whole. Labor demand, however, evolved quite similarly across the country. The comparatively low decline in labor costs observed at the national level is thus, to an important extent, related to South-specific developments and this needs to be borne in mind in designing a wage policy to boost employment. Box 2 examines the labor market problem of the South in more depth.

Interpreting labor supply and demand developments

Taxation

40. Why has there been less of a labor supply increase in Italy, as evidenced by a smaller labor cost decline for given labor market conditions? Was it related to developments intrinsic to labor markets, such as trade unions bidding up wages, or to external factors? The argument developed here is that Italy experienced a sharp increase in the tax burden on labor during the 1980s and 1990s, quite in contrast to the comparator countries. Several points emerge from Table 8: first, Italy had the highest marginal tax wedges on labor income in 1996;¹⁵ second, Italy and, to a much lesser extent, France exhibited a sharp increase in effective tax rates on labor (income taxes and social security payments) over the past two decades, and particularly since 1990 in Italy. This increase in the tax burden made wage moderation much more painful than in the comparator countries.

41. Slow labor cost growth has played an important role in raising employment growth in the Netherlands. Watson and others (1999) underscore that employment growth in this country reflected to a substantial degree agreement among the social partners on the need for slower wage growth, notably following the Wassenaar agreement of November 1982;¹⁶ and they emphasize that wage moderation was supported by reductions of taxes and contributions: according to their evidence, the gross wage of an average production worker rose by about 5 percent in real terms in the decade 1984-93 but the take-home wage by some 10 percent.¹⁷ Developments in Italy were different. As is apparent in Figure 3b, wage moderation did not set in until a decade later than in the Netherlands, namely following the 1992/93 agreement between the social partners—an agreement which was similar to the Wassenaar agreement in some important respects. Since then, real product wages in the business sector have hardly increased; actually, considering the continuing rise in taxation and social security contributions, rough calculations suggest that take-home consumption wages may well have declined in real terms for large segments of the labor force,¹⁸ However, notwithstanding such a potential decline in take-home consumption wages, considering the state of the labor market, gross wages have not displayed a moderation on the scale observed in the Netherlands after 1982, not least owing to tax and contribution hikes. And the effect gross wage moderation on hiring may have been delayed by stringent labor market regulation (see below).

Box 2.

The Mezzogiorno Problem

The most challenging questions facing policymakers in Italy relate to the structural problems of the South. With southern per capita GDP at some 55 percent of the level in the Center-North, standard (neoclassical) growth theory suggests that the South should record considerably higher GDP growth than the Center-North. There is a vast literature on the duality of the Italian economy that blames an array of factors for the underdevelopment of the South: (i) inflexible labor markets, notably trade unions that force a harmonization of labor costs across regions; (ii) rising competition from northern firms as a result of falling transportation costs and a richer southern market—the latter owing to rising public transfers; (iii) a public policy approach that generated a capital-intensive industrial structure, unsuited to support local entrepreneurship; (iv) poor governance; and (v) low human capital.¹ This study focuses on labor costs and employment.

Concerning regional labor markets, the key development has been the sharp decline in employment in the South from 1992 onward (Figure 7). What lies behind the South-specific evolution in employment? Generally, it is the scaling down of government (nonbusiness) employment, as well as the drop in construction activity after the discovery of widespread abuse of public funds (tangentopoli) that has been blamed. In this respect, the figures in Table 14 underscore two points. First, as in the Center-North, the reversal in employment growth in the South after 1992 is largely explained by developments in business services. Second, the turnaround in construction and nonbusiness services in the South accounts for about as much of the South-specific employment growth reversal (i.e., the change in employment growth in the South less the change in the Center- North, both in percent of 1980 total regional employment) as the slowdown in business services. As a result, the difference between regional unemployment rates widened considerably.

The evolutions of labor demand (equation 5) and supply (equation 1) in the South relative to the Center-North are shown in Figure 8. They suggest that labor demand in the South declined by an additional 5 percent. More noticeably, labor supply increased much less than in the Center-North: considering unemployment in the South relative to the Center-North, wages should have declined by some 10 percentage points for β=1 and 15 percentage points for β=2. However, there was little adjustment in practice. This is substantiated further in Brunello and others (2000), who find evidence consistent with aggregate wage setting in Italy depending only on the rate of unemployment prevailing in the Center-North.

The absence of a correlation between actual and predicted relative real wages over time, which has been particularly acute since 1992², probably reflects excessive centralization of wage bargaining, an issue that has been identified as one of the fundamental problems for the South in the literature.

1

On point (i), see for example, Lutz (1963), Bodo and Sestito (1991), Casavola, Gavosto, and Sestito (1995), and Gavosto and Rossi (1997); on (ii), see Graziani (1979) and Faini (1983); on (iii), see Del Monte and Giannola (1978); on (iv) see Triglia (1992) and Tullio and Quarella (1998); and on (v) see Brunetta and Tronti (1994).

2

More recently, the phaseout of social security rebates, to be completed in 2001, on newly-hired employees in the South (mainly in manufacturing) has hampered more relative wage adjustment as measured here (compensation of employees includes employer-paid social security contributions).

View Full Size

Italy: Center-North and South, 1980-99

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: ISTAT; SVIMEZ; and Bank of Italy.

• Download Figure
• Download figure as PowerPoint slide

Table 14.

Italy: Regional Employment Growth, 1980-98

Sources: ISTAT; and SVIMEZ.

Table 14.

Italy: Regional Employment Growth, 1980-98

	Primary	Secondary		Tertiary
		Total	of which: construction	Business	Nonbusiness
	(In percent of total 1980 employment in the region)
Center-North
1980-91	−2.4	−5.6	−0.4	11.2	2.3
1992-98	−1.7	−3.2	−0.2	1.2	−0.1
South
1980-91	−5.9	−1.9	−0.3	12.5	4.6
1992-98	−3.9	−3.4	−1.9	−1.8	−0.5

Sources: ISTAT; and SVIMEZ.

View Table

Table 14.

Italy: Regional Employment Growth, 1980-98

	Primary	Secondary		Tertiary
		Total	of which: construction	Business	Nonbusiness
	(In percent of total 1980 employment in the region)
Center-North
1980-91	−2.4	−5.6	−0.4	11.2	2.3
1992-98	−1.7	−3.2	−0.2	1.2	−0.1
South
1980-91	−5.9	−1.9	−0.3	12.5	4.6
1992-98	−3.9	−3.4	−1.9	−1.8	−0.5

Sources: ISTAT; and SVIMEZ.

View Full Size

Italy: Regional Labor Demand and Supply, 1980-98

Citation: IMF Staff Country Reports 2000, 082; 10.5089/9781451819793.002.A001

Sources: ISTAT; SVIMEZ; and Fund staff calculations.1/ Productivity-adjusted real wage at a constant unemployment rate.

• Download Figure
• Download figure as PowerPoint slide

42. A question that arises is why employment growth was lower in Italy over those years during which effective labor taxation was less high than in most comparator countries. Rising labor taxation can pose more problems than high labor taxation. Assuming that capital is internationally mobile while labor is not, the incidence of labor taxes ultimately falls fully on labor. High labor-taxed countries thus need not have higher unemployment rates. However, following payroll and other tax hikes, unemployment may well rise until unions agree to lower wage floors or cease to seek offsetting wage increases.

43. To understand the relation between labor taxation, wages, and employment, recall that what matters foremost for labor supply is the relation between post-tax pay and the reservation wage. The analysis here has focused on pre-tax pay, including employer-paid social security contributions.¹⁹ Assume, for simplicity, that reservation wages are based on alternative earnings, which are also taxed, and that taxation is proportional. Standard models then suggest that taxation does not matter for employment: the real post-tax wage and the reservation wage drop in tandem.²⁰ However, there are various reasons why post-tax wages may not decline one-for-one as labor taxation rises: (i) individuals’ utility may depend on both the level of income and its change;²¹ (ii) wage floors, such as those in Italy at the sectoral level, can prevent an adjustment in market wages in response to payroll tax hikes; (iii) progressive taxation can narrow differentials between market and reservation wages; and (iv) the reservation wage may be a function of untaxed nonmarket activities, such as tending to children or the elderly, which are important in Italy.²² Under such circumstances, rising labor taxation may result in an increase in pre-tax real wages, and lower employment—a scenario that likely has played an important role in Italy’s labor market developments over the past decades.

44. The extent to which labor taxes affect the relation between post-tax pay and reservation wages both in the short and long run is fundamentally an empirical issue. In this regard the evidence in the literature is ambiguous: Blanchard and Katz (1997), for example, note that the cross-sectional evidence in Europe does not reveal much correlation between tax rates and unemployment rates, nor between changes in tax rates and unemployment rates.

45. By contrast, Daveri and Tabellini (1997) argue that if wages are set by strong trade unions, an increase in labor taxes is shifted onto higher real wages and results in higher unemployment: the continental European countries, unlike the other OECD countries, provide evidence for such a relation within countries over time. The authors’ results suggest that a 10 percentage point increase in effective labor tax rates, through the effect on wages, can account for a 4 percentage point increase in European unemployment. Marino and Rinaldi (2000) find that for OECD countries, an increase in the effective labor tax rate by 1 percentage point (say, from 50 percent to 51 percent) is fully reflected in higher labor casts in the first year.²³ And, for Italy in particular, Brunello and others (2000) observe a cointegrating relationship between unemployment, the tax wedge, the real interest rate, and union power:²⁴ according to their simulations, a reduction in the tax wedge from the level recorded in 1996 to that which prevailed in the early 1980s would yield a 15 percent reduction in the unemployment rate, with about 70 percent of this reduction taking place within five years. In all, although there is no compelling evidence for a cross-sectional relation between taxation levels and unemployment, there is a growing body of research substantiating a time- series relation, that is, whereby an increase in labor taxes may be associated with higher unemployment, particularly for continental Europe.

Union power and skill-biased technological progress

46. What lies behind the declines in labor demand? Two potential explanations are worth highlighting: (i) a reduction in featherbedding related to a decline in union power, including in the context of privatization and restructuring of public enterprises in the business sector; and (ii) the substitution of capital for labor following skill-biased technological progress.

47. While it is difficult to believe that measures to reduce featherbedding can lower labor demand for 20 years, at least for Italy, they may have played a role. First, the demand decline accelerated in the 1990s and thus could be related partly to privatization and public enterprise restructuring. In 1987, nine out of the ten largest firms (by turnover) were in public hands, with their receipts totaling more than the equivalent of 15 percent of GDP. This statistic changed little until 1994, when one of the largest privatization program among OECD countries left only two of these companies state controlled; following privatization, employment losses were sizable. In addition, fiscal consolidation called for restructuring poorly performing national monopolies: since 1990, employment losses in the railway and postal companies alone have amounted to some 160,000, or the equivalent of some ⅔ of 1 percent of the labor force. Second, if featherbedding had played an important role, employment losses would generally have been concentrated in larger firms and these losses would have risen as union power declined: both are documented for Italy in Brunello and others (1999). The drop in union power, which started in the 1980s, is documented in further detail in Bertola and Ichino (1995). And third, there is some evidence that actual per capita working hours have increased, while contractual hours have declined slightly: Casadio and D’Aurizio (1999) document that per capita working hours in manufacturing firms with more than 50 employees rose 12 percent during 1985-98, despite a small reduction in contractual hours. They observed that more cooperative industrial relations, as evidenced by structural reductions in labor disputes and absenteeism, played an important role: both are equivalent to less featherbedding.²⁵

48. Skill-biased technological progress can be thought of as a mean-preserving increase in the spread of the productivity distribution for workers across skill categories. If wage differentials do not widen in response (e.g., because unemployment benefits put a floor on reservation wages or due to egalitarian wage setting), such progress raises unemployment and may prompt substitution of capital for low-skilled labor.²⁶

49. Within professions, wage differentials are currently lowest in Italy (Table 15), although across professions the evidence is less clear. Investigating the issue of wage differentiation more formally for the 1980s, Freeman and Katz (1995) find that the United Kingdom experienced widening differentials similar to those of the United States; however, France, Italy, and the Netherlands did not. Erickson and Ichino (1995) show that Italy experienced a strong compression of wage differentials during the 1970s, similar to the better-known situation in Sweden. While this compression came to a stop around 1982-83, there was no obvious reopening of differentials through 1990: by contrast, their results suggest that in the United States not only was the compensation structure less compressed, it also widened considerably. Brandolini and Sestito (1999) revisit and update the work of Erickson and Ichino: they find similar evidence for Italy for the 1980s but observe a considerable widening of wage differentials during 1990-95; for male workers the widening even goes somewhat beyond levels recorded in the late 1970s. Their evidence is, therefore, not supportive of a labor-demand decline during the 1990s driven by insufficient wage differentiation, unless it is interpreted as a continued response to developments in earlier decades, where it may have played an important role.

Table 15.

Italy: Earnings Differentials, 1995

Source: EUROSTAT, Structure of Earnings Statistics, 1995.

^1/Standard deviation of average gross hourly earnings of listed professions divided by the average, in industry and markets services.

Table 15.

Italy: Earnings Differentials, 1995

		France	Italy	Netherlands	United Kingdom
Variance across professions 1/		47.1	62.2	23.7	36.9
Variance within professions: 1/
	Corporate managers	67.5	42.6	64.2	71.2
	Managers of small enterprises	52.0	39.1	55.4	69.3
	Professionals	50.3	39.5	53.4	48.6
	Technicians and associate professionals	54.3	45.5	48.3	79.1
	Clerks	43.2	32.0	38.4	35.0
	Sales agents	44.7	36.2	45.7	46.0
	Craft and related workers	45.8	25.4	46.0	33.2
	Plant and machine operators and assemblers	39.4	37.2	38.5	36.9
	Elementary occupations	36.2	25.4	55.3	42.9

Source: EUROSTAT, Structure of Earnings Statistics, 1995.

^1/Standard deviation of average gross hourly earnings of listed professions divided by the average, in industry and markets services.

View Table

Table 15.

Italy: Earnings Differentials, 1995

		France	Italy	Netherlands	United Kingdom
Variance across professions 1/		47.1	62.2	23.7	36.9
Variance within professions: 1/
	Corporate managers	67.5	42.6	64.2	71.2
	Managers of small enterprises	52.0	39.1	55.4	69.3
	Professionals	50.3	39.5	53.4	48.6
	Technicians and associate professionals	54.3	45.5	48.3	79.1
	Clerks	43.2	32.0	38.4	35.0
	Sales agents	44.7	36.2	45.7	46.0
	Craft and related workers	45.8	25.4	46.0	33.2
	Plant and machine operators and assemblers	39.4	37.2	38.5	36.9
	Elementary occupations	36.2	25.4	55.3	42.9

Source: EUROSTAT, Structure of Earnings Statistics, 1995.

^1/Standard deviation of average gross hourly earnings of listed professions divided by the average, in industry and markets services.

Unemployment benefits and labor market regulations

50. What about the role of unemployment benefits and labor market regulation? Benefits affect the supply of labor by setting a floor on reservation wages. Their variation across countries offers little in explaining differential labor supply (Table 10b).²⁷

51. Hiring and firing restrictions reduce both employment inflows and outflows, leaving an ambiguous effect on the unemployment rate.²⁸ However, by virtue of making employers more reluctant to fire as well as hire such legislation could explain the delayed response of employment to wage moderation in France and Italy: from the point of view of regulation, both countries compare unfavorably against the other countries in the sample (Table 10a).²⁹ This argument is developed further in Bertola and Ichino (1995): they argue that the wage moderation during 1993-95 in Italy may not have been sufficient to overcome employers’ reluctance to hire, considering the difficulties involved in firing permanent employees. It is noteworthy that only following the reforms of fixed-term employment and temporary work agencies in the second half of the 1990s that employment grew rapidly, considering the low real GDP growth rates.

52. Turning to part-time employment, regulations were stricter in Italy and France. This could have adversely affected labor supply, particularly of women, which has boomed in the Netherlands.³⁰ Importantly, flexible employment can raise firms’ demand for labor by enabling them to fine-tune working time and output and thereby meet orders “just in time.”³¹ A more liberal handling of such contracts, combined with wage moderation, appears to have an important effect on labor market developments in Italy.