Euro Area Policies: Selected Issues
Author:
International Monetary Fund
Search for other papers by International Monetary Fund in
Current site
Google Scholar
Close

The paper discusses the purpose, properties, and theoretical foundations of various indicators of inflation and also describes the forecasting methodology and performance of these indicators. It reviews the successful European labor market reform experiences and analyzes regulatory and supervisory frameworks in the European Union (EU), and assessments carried out under the IMF-World Bank Financial Sector Assessment Program (FSAP). It also investigates whether the cross-country correlation of bank business in Europe makes a good case for pan-European supervision.

Abstract

The paper discusses the purpose, properties, and theoretical foundations of various indicators of inflation and also describes the forecasting methodology and performance of these indicators. It reviews the successful European labor market reform experiences and analyzes regulatory and supervisory frameworks in the European Union (EU), and assessments carried out under the IMF-World Bank Financial Sector Assessment Program (FSAP). It also investigates whether the cross-country correlation of bank business in Europe makes a good case for pan-European supervision.

II. Lessons from Successful Labor Market Reformers in Europe17

A. Introduction

27. A vigorous debate is underway in Europe over the future of economic reform and equity. Arising from its history, including the massive upheavals during the first part of the twentieth century, Europe has developed a solid attachment to the principal of solidarity. But notwithstanding some reforms and progress, unemployment remains persistently high and per capita income has been stuck at about three-quarters of the United States level since the mid-1970s. More recently, productivity also suffered a protracted slowdown, as countries found it difficult to adapt to ongoing technological change and globalization. The upshot is stalled growth, with limited prospects for improvement in light of the impending onset of population aging. In the unfolding debate, many point the finger at Europe’s welfare states, arguing that high levels of taxes and transfers hurt employment, while underlying rigidities hinder an effective reallocation of resources.

28. Some argue that efficiency and equity are by no means incompatible. Sapir (2005), for instance, points to the “Nordic” countries as having the best of both worlds, combining high employment with high degrees of equity. In contrast, the “Mediterranean” countries—which rely more on employment protection legislation (EPL) than income support—score poorly on both counts. Others lie in the middle, with the “Anglo-Saxon” countries scoring high on efficiency and low on equity, while the opposite is true for “Continental” countries characterized by generous unemployment benefits and strong unions.

29. Rather than enter the “social models” debate directly, this chapter focuses on successful European reform experiences. It singles out four countries in particular—Denmark, Ireland, the Netherlands, and the United Kingdom—that have achieved some degree of success in reducing unemployment and stimulating job growth, despite vastly differing economic conditions. To complement the case study analysis, empirical techniques (event studies and econometric analysis) will also be utilized. In this process, all aspects of reform will be considered including background, political economy context, fiscal policy behavior, the role of labor market institutions, and patterns of poverty and inequality. The goal is to lay out a set of basic conclusions about what reform strategies work in the European context that can be used as a guide for the future.

30. The four case study countries stand out in terms of their success in reducing unemployment and increasing employment (see Table II.1). Four of the lowest five unemployment rates can be found in these countries, and they also covered the most ground over two decades (Denmark slightly less than others, but only because its peak unemployment was in the early 1990s, not the early 1980s). Broadening the picture to look at the overall employment rate conveys similar information, with Ireland and the Netherlands leading the pack (Figures II.1 and II.2); the same holds true for both male and female employment (Figures II.3 and II.4). While all case study countries succeeded in lowering unemployment, Ireland and the Netherlands took further advantage of female participation catch-up18. Ireland also benefited from uniquely favorable demographics (Table II.2).19 But the gains in male employment were effectively neutralized by declining male participation, a pattern also evident in Denmark and the United Kingdom.

Table II.1:

Unemployment in the EU-15

article image
Source: OECD
Figure II.1.
Figure II.1.

Employment Growth Since 1980

(1980=100)

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.
Figure II.2.
Figure II.2.

Breakdown of Employment

(Total)

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.
Figure II.3.
Figure II.3.

Breakdown of Employment

(Men)

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.
Figure II.4.
Figure II.4.

Breakdown of Employment

(Women)

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.
Table II.2.

Breakdown of Total Employment Growth in European Union Across Two Decades

article image
Source: OECD

31. Adjusting for government employment and hours worked reveals some differences, especially for Denmark. Stripping the government component out of total employment does not change the story for Ireland or the Netherlands, as employment gains in these countries came unambiguously from the private sector (Figure II.1). In terms of the other countries, the United Kingdom looks somewhat better, and Denmark somewhat worse.20 Although part-time prevalence increased, especially in the Netherlands (Figure II.5), total hours in the business sector nonetheless grew robustly (Figure II.1). Still, in relative terms, the Netherlands slips behind the United Kingdom, while Denmark’s performance was average, recording practically no growth in hours since 1980.

Figure II.5.
Figure II.5.

Prevalence of Part-time Employment

(in percent of total employment)

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.

B. Crisis and Reform: A Cursory Overview

The setting

32. In the face of an adverse global supply shock at the start of the 1980s, the case study countries all made policy mistakes. Some countries, including Ireland and Denmark, responded inappropriately by engaging in an expansionary demand policy in the face of a supply shock. Wages spiraled out of control in some countries, most starkly in the Netherlands, where the wage share hit 95 percent, 12 percentage points higher than a decade earlier (Bakker, 1999). In the United Kingdom, unit labor costs also expanded throughout the 1970s, as workers gained large wage increases in the face of persistent low productivity growth, and unions struck a particularly belligerent attitude. The Netherlands also engaged in an overly-rapid increase in the expenditures, prompted partly rising natural gas revenues.21

33. By the early 1980s, as a result, all four countries faced severe macroeconomic crises. Denmark, the Netherlands, and the United Kingdom all faced two consecutive years of negative growth, with real GDP declining by 2½ percent, 2 percent, and 3½ percent respectively. Inflation rose across the board, hovering around 18–20 percent in Ireland and the United Kingdom, and reaching 10–12 percent in Denmark, and 7 percent in the Netherlands. Likewise, unemployment trended upwards, reaching 8–10 percent in Denmark, the Netherlands, and the United Kingdom, and surpassing 16 percent in Ireland.

34. Not surprisingly, public finances also spiraled out of control, at least in three of the four countries. Underlying imbalances combined with rising interest rates in the early 1980s and led quickly to unstable debt dynamics. Large fiscal deficits emerged in Denmark (8½ percent of GDP), Ireland (13 percent of GDP), and the Netherlands (6 percent of GDP). But public finances in the United Kingdom were relatively well behaved: the general government deficit averaged 3¾ percent of GDP between 1979–81, no major change over the previous five years.

The reform strategy

35. In response to the crises, the case study countries all leveraged the complementarities between fiscal consolidation and reforms to enhance labor supply (see text table for a basic summary of reforms). Ireland and the Netherlands centered their reforms on consensus-based agreements between social partners, trading wage moderation for labor tax cuts. Instead, the United Kingdom weakened the power of unions. Rather than address union behavior directly, Denmark concentrated on benefits reform, by combining continued generosity with lower duration, tougher conditionality, and stricter activation requirements. Most countries addressed benefits reform to varying degrees. All four countries also reduced tax wedges on labor income. Furthermore, Denmark and the Netherlands boosted active labor market policies (ALMPs). These countries also loosened restrictions on temporary (but not permanent) employment protection legislation (EPL). However, the case study countries tended to be among the most liberal in terms of labor market restrictions. Also, all liberalized product markets over this period. Another key point is that all four countries engaged in complementary fiscal adjustment, often substantial. And three of the four countries set the stage for a long-term decline in the size of government (except Denmark).

Text Table.

Matrix of Reforms

article image
Source: Author’s assessment.

= significant reform;

= moderate reform.

No reforms but low to start with.

36. While crisis spurred reforms in all countries, the continuation of reforms depended on favorable macroeconomic circumstances. The case study countries all experienced growth and employment booms following the advent of the labor market reform phase. And the employment booms were not achieved at the expense of labor productivity. In Ireland and the Netherlands in particular, reforms continued over an extended period of time, and even intensified, as growth remained solid. But in Denmark, unit labor costs increased on the back of large wage increases in 1987, and growth came to a temporary standstill, giving rise to a pause in the reform agenda. In the United Kingdom, reforms took some time to bear fruit as unemployment remained elevated for almost a decade, perhaps due to inadequate monetary policy credibility (Pissarides, 2003). Over the longer horizon, however, the reform momentum was maintained in all countries.

C. Labor Market Reform

37. At the cornerstone of reforms in all countries was a desire to boost labor supply. Based on a model whereby unions and employers bargain over wages, “wage moderation” can be interpreted as structural changes in unions’ approach to wage bargaining, or an outward shift in the labor supply, or wage, curve.22 The wage variable used in this paper is the productivity- and cyclically-adjusted real hourly compensation rate.23 A number of factors can lead to outward shifts in the wage curve: (i) changes in the attitudes of unions and workers, placing a greater emphasis on employment; (ii) falling labor taxation, allowing workers to accept lower gross wages for the same net wage; (iii) unemployment benefit reform, reducing the reservation utility of union members; (iv) reducing government employment or government wages, also reducing reservation utility, given that government employment is an alternative to private employment (see Ardagna, 2004). The reforming countries did indeed place structural increases in labor supply high on the agenda (Figure II.8 and Table II.3). Looking across the two decade horizon, the countries exhibiting the biggest shifts in labor supply are Ireland, Finland, Sweden, the United Kingdom, and the Netherlands. More countries joined them in the late 1990s and early 2000s, including Italy, Spain, and Germany.

Figure II.6.
Figure II.6.

Fiscal Developments I

(in percent of GDP, unless otherwise noted)

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.
Figure II.7.
Figure II.7.

Fiscal Developments II

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD.
Figure II.8.
Figure II.8.

Real Wages, Tax Wedge, and Replacement Rate

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD
Table II.3.

Periods of Wage Moderation in Case Study Countries, 1980–2003 1/

(cumulative percentage change over the period)

article image
Source: Author’s calculation from OECD data.

Real productivity and unemployment-adjusted hourly compensation.

Behavior of unions

38. Wage moderation was abetted in Ireland and the Netherlands by coordinated agreements between social partners, while the United Kingdom focused on reducing the power of unions. The Netherlands enjoyed a long history of corporatism and consensus-driven policymaking, especially in the domain of wage bargaining (Hartog, 1999). Moreover, certain aspects of the legal institutional framework fostered the consensus-based approach. Laws from the interwar period made any agreement with a union binding on all the workers in the firm, and allowed the government to extend a collective agreement that covered a large majority of the industry to the entire industry. Ireland, on the other hand, had no such corporatist history, but unions had a vested interest in aligning themselves with this new approach, given the gravity of the crisis and the fact that union membership was waning.24 Some believed that such a strategy would have been impossible in the decentralized bargaining system in the United Kingdom (Nickell and van Ours, 1999).25 But given the Irish institutional transformation, perhaps its size and heterogeneity, as well as the greater ideological polarization, played a greater role in hindering a consensus-based approach from emerging in the United Kingdom.

39. At the core of the Dutch and Irish programs was the strategy to mitigate the effects of lower nominal wage growth with labor tax cuts. The seminal Wassenaar agreement in the Netherlands was signed in 1982 between the leading labor federation and the employer’s federation, trading wage restraint for working hour reductions, alongside government commitments to reduce labor taxes and social security contributions. Unions also ceased their opposition to part-time jobs. While the gross real wage of the average worker increased by less than 1 percent between 1983–98, the corresponding real net wage rose by nearly 15 percent (Bakker and Halikias, 1999). Ireland adopted a similar strategy in 1987, without the emphasis on working hours. Although the wage agreements applied only explicitly to the unionized sector, the agreements acted as a more general benchmark for wage expectations. Given its success, this strategy has continued unabated in both countries. The Wassenaar agreement was updated in 1989, 1993, and 1997, and a total of six medium-term agreements were negotiated in Ireland.26 As the fiscal accounts improved, tax cuts became more feasible.

40. In the United Kingdom, union power was reduced progressively throughout the 1980s. Statutory recognition procedures were abolished, actions that force contracts with union employers were prohibited, the grounds for refusing to join a union were extended, picketing was limited, union immunities were weakened, and pre-strike ballots were required. Significantly, the closed shop was banned, allowing employers to hire non-union workers. Many firms chose not to recognize unions at workplaces at all, and others insisted on plant-specific agreements (Blanchflower and Freeman, 1993). With the exception of the public sector, unions were concentrated in the older manufacturing sector, the very sector which experienced the greatest decline over this period. Overall, union density fell from 50 percent in 1980 to 30 percent by end of the 1990s. More revealingly, union coverage, the percentage of employees subject to a union agreement, fell from 70 percent in 1980 to only 35 percent at this time.

41. Unlike in the other cases, a shift in wage bargaining behavior was not a key aspect of the Danish reforms. As in the United Kingdom, the Danish government initially adopted a confrontational approach with unions, which led to major industrial unrest and ultimately failed to restrain wage growth. Following this, the wage bargaining system became increasingly decentralized over the 1990s, as employers largely abandoned coordinated bargaining, and the government switched direction and focused instead on reforming labor market institutions (see next page).

Labor market institutions27

42. To boost labor supply, the reforming countries all engaged in institutional reform, but focused on different aspects. To varying degrees, the four case study countries cut tax wedges, reformed the benefits regime, made recourse to ALMPs, and reduced EPLs.

43. A common factor among the four countries was the steady reduction in tax wedges over the period in question. The overall tax wedge on labor income28 declined markedly over two decades in Ireland, the Netherlands, and the United Kingdom and modestly in Denmark (Figure II.8). This was not a common trend, as the wedge actually rose in half of the EU countries at this time. Also, while Ireland and the United Kingdom are positioned on the lower end, Denmark and the Netherlands are around average, lower than countries like France, Germany, Italy, Belgium, Finland, and Sweden. In Ireland, the Netherlands, and the United Kingdom, the decline in the tax wedge represented a clear government policy of compensating for wage moderation by cutting taxes on labor income.

44. The case study countries all engaged in some aspect of benefit reform, reducing the level of unemployment benefits, or their duration, or strengthening eligibility requirements. In this context, Figure II.8 shows that the gross replacement rate in the first year of unemployment declined in Denmark, Ireland, and the United Kingdom over the relevant period.29 Reforms led to less generous benefits in the United Kingdom and the Netherlands in particular (the latter focused on sickness and disability as well as unemployment), while in Ireland, benefits failed to keep pace with after-pay income. But, despite changes, the replacement rate in Denmark and the Netherlands remained high, near the top in the EU, while Ireland and the United Kingdom sit on the other end of the scale. But lessening the replacement rate was only one dimension of reforms. Denmark and the Netherlands also cut the maximum duration of unemployment benefits, while three of the four countries (bar Ireland) tightened up eligibility requirements for receipt of benefits. In Denmark, policies deliberately allowed for generous benefits in the face of shorter duration and tougher eligibility conditions. Indeed, Grubb (2000) noted that a notable common factor among reformers was a stricter enforcement of job search and better surveillance of eligibility.

45. In the 1990s, the Netherlands also shifted toward the “privatization” of social security, as risks were shifted to employers and private insurance schemes. The whole system was decentralized, minimizing the role of social partners. The old system created some poor incentives, as all parties—workers, employers, and the governments—had their own motives for preferring the disability to the unemployment benefit channel (Barrell and Genre, 1999). From 1994, employers were made responsible for the first six weeks of sick leave, and this was extended to a full year in 1996, after which the person moved to disability status. From 1998, disability premia were fully paid by employers, and differentiated by enterprise. From 1997, management of the social security system, until then in the hands of the unions and employers organizations—was contracted out to social security agencies, which were allowed to compete with each other, and with private companies (see Bakker and Halikias, 1999).

46. The case study countries relied on ALMPs to differing degrees. Active measures are prominent in Denmark and the Netherlands, although Ireland is also above average in terms of expenditure. They are not significant in the United Kingdom. And while ALMPs grew in prominence in both Denmark and the Netherlands over the reform period, their popularity actually declined in Ireland and the United Kingdom (Figure II.9). In Denmark, the unemployed were required to take part in ALMPs, and requirements on this front became progressively tougher throughout the 1990s. Countries also differed in their choice of ALMPs (Figure II.10), with Ireland and the Netherlands focusing on subsidized employment, which some argue is most effective (Estevao, 2003). ALMPs combined with activation (as introduced in Denmark and the United Kingdom) may also work well (Martin and Grubb, 2001). Indeed, youth unemployment fell by 40 percent in the United Kingdom after the onset of the “New Deal” (Swagel, 2000). Likewise, the youth unemployment rate in Denmark fell from a peak of 16 percent to 3 percent (Andersen, 2003).

Figure II.9.
Figure II.9.

Labor Market Institutions

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD
Figure II.10.
Figure II.10.

Composition of Active Labor Market Programs

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD

47. A further common element is that the case study countries do not rely heavily on employment protection. Indeed, in the early 2000s, the four countries clustered around the bottom of the group in terms of EPLs, with highly liberalized labor markets in Ireland and the United Kingdom in particular (Figure II.9). EPLs related to temporary employment were also lowered in Denmark and the Netherlands over the reform period. But while many countries followed this liberalization pattern, there was little movement in loosening regulations on permanent employment (OECD, 2004). In Denmark, a liberal approach to regulation is part of its much-touted “flexicurity” model, which insures workers against income loss, but not job loss. Indeed, job turnover among Danish workers is around 30 percent, matched only by the United Kingdom. The percent of laid off workers re-hired by the same company is also high; 40 percent in 1998 were re-hired within six months (Gaard and Kieler, 2004).

Product market regulation

48. Product market liberalization can be a boon to labor market reforms. Empirically, product market liberalization has been shown to spur labor market liberalization (Debrun and Annett, 2004; OECD, 2006). Estevao (2005) finds that excessive regulation suppresses the beneficial effects of labor market reform on employment and output, by inhibiting competition or discouraging entry into the market. Likewise, Blanchard and Giavazzi (2003) argue that product market reforms lower prices and increase real wages, thus setting the groundwork for labor market reforms. Berger and Danninger (2005) also show that labor and product market reforms are complementary.

49. The case study countries tend to favor less stringent product market regulation. Indicators of product market regulation are imperfect. The most broad-based indicators, encompassing the whole economy, are only available from the OECD for 1998 and 2003. These data suggest that the four case study countries are the least regulated in the sample in 1998, and in 2003, the Netherlands slips to fifth place behind Sweden. A recent European Commission study showed that firms in the United Kingdom tended to have the lowest cost and waiting time in terms of starting a new business (Maher and Wise, 2005). The only available time series conveys information on seven non-manufacturing sectors—airlines, telecoms, electricity, gas, post, rail, and road freight—between 1975–2003 (see Conway and Nicoletti, 2006). Here, the picture is not as clearcut. While the United Kingdom, Netherlands, and Denmark are still at the bottom, Ireland is now above average, largely due to restrictions in railways, gas, and airlines.30 In terms of the dynamic pattern, again, Denmark, the Netherlands, and the United Kingdom liberalized most over a twenty-year period, but the trend has been downward for all countries (Figure II.9).

50. Deregulation and privatization were prominent features of reform. Denmark and the Netherlands began liberalizing utilities in the early 1990s. Also in the Netherlands, competition regulation was loosened, and shop opening hours were extended in 1996 (Barrell and Genre, 1999). The United Kingdom engaged heavily in privatization: 2 percent of GDP was produced by public companies in 1997, down from 12 percent in 1979, and around 6 percent of the workforce shifted from the public to the private sector over twenty years (Card and Freeman, 2002).

D. Fiscal Policy Developments

51. Fiscal and structural reforms went in one direction, as the case study countries undertook extensive fiscal adjustment in unison with labor supply shifts (see Figures II.6 and II.7). From the perspective of two decades, the fiscal turnaround in Denmark, Ireland, and the Netherlands in particular has been impressive.

  • All four case study countries undertook substantial adjustment during various phases over the past two decades. Leading the pack, the cyclically-adjusted primary balance (CAPB) in Denmark improved by a full 13 percentage points from 1982–86. Following some slippage, Denmark did not see sustained consolidation again until the late 1990s. Ireland adjusted by about 6½ percentage points from 1986–89. Adjustment over the 1990s was more modest, but fiscal policy remained prudent. Consolidation in the Netherlands also proceeded in two distinct phases, by nearly 4 percentage points over the early 1980s and again by 5½ percentage points in the early 1990s. The United Kingdom followed a similar pattern, with an early CAPB improvement of almost 5 percentage points between 1979–82, and again between 1994–99 (7½ percentage points).

  • The most successful episodes tended to be expenditure-based (see Alesina and Ardagna, 1998). In Ireland, the stabilization program included a freeze on the public sector wage bill, as well as cutbacks in capital spending. In two years alone, public employment fell by more than 10 percent. The Netherlands embarked on a similar path, as the government wage bill was held down both by containing salaries (including through a 3 percent nominal cut in 1983) and employment. Social benefits were also cut, with a nominal freeze in the minimum wage and minimum benefit and reduced replacement rates. The inauguration of the framework underpinned by expenditure rules in 1994 meant that fiscal discipline became entrenched. The mid-1990s adjustments in Denmark and the United Kingdom were also expenditure-based, with transfers falling in both, and also government wages declining in the latter.

  • Countries relying on labor tax-based consolidation tended to run into difficulties. Although Denmark cut expenditure substantially during its first adjustment (1982-86), the fact that it relied equally on labor tax increases eventually contributed to wage pressures in 1987 and an end to the accompanying growth boom. Similarly, in Ireland, an earlier consolidation episode (1981–84) based prominently on labor taxes was not successful. Although the United Kingdom’s early adjustment was also based on tax increases, the aim of the government was to shift away from labor taxes toward VAT.

  • Many of the most successful episodes twinned expenditure and labor tax cuts. This was the case during both Dutch adjustment episodes, Ireland in the late 1980s, and Denmark in the late 1990s. But restoration of fiscal prudence preceded tax cuts in all cases.

52. Taking a longer-term perspective, three of the four countries (Ireland, the Netherlands, and the United Kingdom) reduced the size of government over two decades. Some of the changes are stark. Ireland, the Netherlands, and the United Kingdom reduced their expenditure ratios by 19, 11, and 7 percentage points respectively. The same pattern emerges from the crucial social expenditure category. Aside from Denmark, all countries also cut government employment, which fell as a ratio of total employment. These diverging trends and the relative rankings carry over to the revenue ratio, which fell by nearly 7 percentage points in Ireland and the Netherlands, and by 4 percentage points in the United Kingdom. Most other countries actually increased revenue, including those that had started from a low base (Greece, Spain, and Portugal). But the size of government in Denmark barely budged and remains the second largest in the EU. In contrast, Ireland and the United Kingdom have the smallest and third smallest governments respectively, while the Netherlands has settled at close to average. Cyclically-adjusted aggregates tell the same story (Figure II.7).

53. Despite diverging opinions, there is good reason to think that fiscal adjustment and labor supply reforms are complementary, especially when the adjustment is expenditure-based. Some have argued that, with a fixed amount of political capital, fiscal adjustment tends to retard structural adjustment. Both IMF (2004) and OECD (2006) argue that while a strong budget balance increases the chances of labor market reform, consolidation hinders it. But, as seen from the case study evidence, fiscal adjustment and reforms that address labor supply may go hand-in-hand. Expansionary fiscal contractions tend to be based on expenditure cuts (especially on government employment, government wages, and transfers) and are accompanied by wage moderation (Alesina and Ardagna, 1998). The labor market channel is clear, as reducing both government wages and transfers, as well as cutting labor taxes, prompts unions to accept lower wages, which in turn leads to higher profitability and—particularly if product markets are liberal—higher employment and growth (Ardagna, 2004; Alesina and others, 2002). Granting tax cuts also neuters potential opposition to labor supply reforms.

54. Others have also shown that tapping the synergies between expenditure adjustment and other reforms in a comprehensive package can be fruitful. Hauptmeier, Heipertz, and Schuknecht (2006), also adopting a case study approach, find that countries typically enjoyed high growth and employment in the wake of substantial expenditure reform programs centered on cutting transfers and the government wage bill, even though they began the retrenchment in the context of recessions or crises.31 Especially in the countries that started early, the fiscal adjustment was typically cast in the context of a broader package, focusing on such factors as macroeconomic stabilization, privatization, and labor market reforms (including liberalization, and reforming tax and benefits systems). To support the expenditure restraint, budgetary institutions were also reformed in numerous countries. Developments in this regard included the use of expenditure ceilings and other fiscal rules, fiscal contracts in the context of coalition governments, multi-year budgeting, legal and procedural reforms, strengthening the role of the minister of finance, and utilizing independent fiscal councils (see also Hallerberg, 2004; European Commission, 2006).

E. Political Economy of Reform

55. The case study experiences show that governments can overcome the well-known political economy obstacles to reform if they stay committed to a consistent set of policies in the labor supply and fiscal arenas. It is well established that the opponents of reform, even if a minority, may be more visible and well organized than the disparate gainers. Even reforms that benefit nearly everybody may still be rejected by the majority if there is uncertainty about who gains and who loses (Fernandez and Rodrik, 1991).

Adjustment could also be delayed by a “war of attrition” between different groups, as each side holds out in the hope that the other side will cave and bear the full costs of adjustment (Alesina and Drazen, 1991). But the reform experiences in the case study countries show how these distortions can be surmounted. More particularly, the following lessons can be gleaned:

  • Wile crisis leads to reforms, the continuation of the reform effort depends on favorable economic conditions. Certainly crisis spurred reform in all four countries. In terms of average growth performance between 1980–82, three of the countries—Denmark, the Netherlands, and the United Kingdom—were among the worst four performers (Greece was the other). And Ireland chalked up the second worst average fiscal deficit, and the third worst average inflation performance, at this time. But successful reforms heralded strong growth, and provided space for the reforms to continue. It is therefore key that sound economic conditions be laid in place, including a consistent and coherent policy mix, without backtracking. This is consistent with the prevailing literature. IMF (2004), for example, finds that reforms are more likely to take place following a sequence of “bad” years. But OECD (2006) also shows that while reforms are usually triggered by crises, some labor market reforms—those related to tax wedges, EPLs, and benefits systems—are most likely to occur during upswings. Certainly, reform of labor market institutions may engender less opposition in good times (Blanchard, 2005).

  • No single type of political system is more suited to reforms than others. The countries in the sample have experienced the gamut of governments. Denmark has a history of minority coalition governments. Ireland has oscillated between single party and majority coalition governments. With a single national electoral district, the Netherlands has the most proportional electoral system in Europe, and multiparty majority coalitions are the norm. And with a majoritarian electoral system, single-party majority governments are standard in the United Kingdom. Many have noted that, while strong stable governments can push through a reform agenda with vigor, this does not preclude weaker governments from seizing the reform mantle, particularly if they allow different groups to “buy into” the process (Castanheira and others, 2006). Indeed, the case study countries went down this route to varying degrees as evidenced by the trade-offs between wage moderation and tax cuts in Ireland and the Netherlands, and the Danish policy of guaranteeing high benefits in return for a tougher system.

  • In all cases, reforms were implemented by a new government, in a decisive break with the past. In Denmark, the adjustment program was begun by a newly-elected government (the “four-leafed clover”) in 1982 that for the first time excluded the Social Democrats. The return of the Social Democrats after a long absence in 1994 heralded the second phase of reforms. In Ireland, a newly-elected single-party minority government laid the groundwork for successful adjustment starting in 1987. Likewise, in the Netherlands, a newly-elected coalition oversaw the Wassenaar agreement, and, for the first time, set budget targets for the life of the government. It was also a new government in 1994—excluding the Christian Democrats for the first time in the postwar period—that launched the expenditure rules. The United Kingdom also had a fresh start under the new Conservative government in 1979.

  • Governments tended to win re-election following reforms. Once the economic preconditions are in place, the electoral process rewards politicians for success. In Denmark, the government was forced to call elections following a budgetary defeat in 1984, as the opposition parties refused to back benefit cuts. Following an election victory, it continued the adjustment program. The Irish, Dutch, and British governments also secured reelection.

  • Political instability is a bane to reforms. The pre-reform period in Ireland was marked by instability, with short-lived governments and deadlocked coalitions. It was not until the major opposition party in 1987 pledged that it would not oppose the government’s economic policy that the reform program began in earnest. In contrast, the single-party government in the United Kingdom maintained effective intra-party discipline throughout the 1980s, despite a deep recession, while facing a weak and divided opposition. Governments in the Netherlands found it useful to put together detailed coalition agreements in advance, typically centering on fiscal policy goals, to avoid any later conflicts. Ireland later adopted this practice with considerable success. Also in Denmark, institutional reforms reduced the ability of parliament to attack individual budgetary measures (Hallerberg, 2004).

  • Once a reform program is successful, it tends to be emulated by future governments, even if of a different ideological hue. In Ireland and the Netherlands, the basic principals of social partnership, centered on wage moderation and fiscal policy commitments, became a mainstay of politics. In the United Kingdom, the new Labour party government in 1997 did not mark a major economic policy break with the past.

  • Independent non-partisan bodies can help shape the public debate on the need for reform. In the Netherlands, the changing mindset was influenced by research from the Central Planning Bureau (CPB), pointing out problems with the Keynesian approach that was dominant in the 1970s.

Finally, given the noted overlap between fiscal adjustment and labor supply reforms, it is worth noting that many of the same political economy factors—including existence of crises, the election of a new government, and strong stable governments—are also seen as factors that influence the former (Alesina, Ardagna, and Trebbi, 2006).

F. The Social Dimension

56. First and foremost, the twin goals of efficiency and equity are not incompatible. (Figure II.11). As noted at the outset, a number of recent commentators have pointed out that high employment can go hand-in-hand with equity (Sapir, 2005; De Groot, Nahuis, and Tang, 2004). In this regard, countries like Denmark and the Netherlands have consistently outperformed other European countries when it comes to indicators like inequality and poverty. Gini coefficients—which measure the degree of inequality in terms of household disposable income—show Denmark as the most equal country in Europe, with the Netherlands not far behind. And although the United Kingdom and Ireland are located at the other end of the scale, they still score better than countries like Greece, Italy, Portugal, or Spain. However, the negative correlation between employment and inequality is less clear for non-government employment, as many of the Nordic countries in particular have preferences for large public sectors. In terms of poverty, Denmark again has the lowest poverty rate, with the Netherlands holding third place after Sweden; in this case however, Ireland sits at the bottom. Also, the degree of equity a country seems related to the size of social expenditure.

Figure II.11.
Figure II.11.

Social Developments

Citation: IMF Staff Country Reports 2006, 288; 10.5089/9781451813050.002.A002

Source: OECD

57. From a dynamic perspective, reform need not require social cohesion to be sacrificed. Denmark and the Netherlands remained in favorable positions after their reforms. And while inequality nudged up in the United Kingdom following the onset of reforms, it was stable in Ireland throughout the adjustment period, and even declined a little (see Callan and Nolan, 1999). This is especially notable given that inequality seems to be related to social expenditure, which fell significantly in some of the reformers—Ireland and the Netherlands in particular. Also, the uptick in inequality over the 1990s was a widespread phenomenon, common to some reformers and non-reformers alike. Some have pointed out that unfavorable effects of expenditure reform on income distribution tend to be mitigated by faster growth and better targeting of public spending (Tanzi and Schuknecht, 2005). And although poverty rates increased in three of the case study countries (except for Denmark), the relative positions of the countries remained unchanged.

58. Social problems persist in all countries, however, including the most equitable ones. Poverty rates remain elevated in Ireland and the United Kingdom. And despite the success of reforms in generating employment, labor inactivity remains rampant in some countries. Indeed, disability expenditure still hovers around 4 percent of GDP in Denmark and the Netherlands, even though it fell by 3 percentage points in the latter. But while Dutch reforms successfully tackled sick leave, disability leave still accounts for about 13 percent of employment (Detragiache, 2002). Likewise in Denmark, more than 20 percent of the working age population continued to receive benefits from all leave and unemployment schemes (Gaard and Kieler, 2004). Integration of immigrants also remains low in Denmark, with participation rates 20 percentage points lower (for men), and 40 percentage points lower (for women) than for native-born Danes (Andersen, 2003). The United Kingdom has a problem with the concentration of inactivity within jobless households, with 17 percent of households with no adult working by the late 1990s (Swagel, 2000). And educational disparities remained a problem: an OECD study in 1997 showed that 22 percent of the working age population in the United Kingdom was at the lowest level of literacy, more than twice the rate in other northern European countries.

G. Beyond the Case Studies

59. To complement and broaden the scope of the case-study analysis, this final section introduces a quantitative dimension, with a simple event study and some basic econometric analysis. Rather than focus on a group of predetermined countries, the event study methodology essentially allows the data to decide. Of course, it is centered on a particular episode in time, and cannot capture whether reforms are sustained over a longer horizon. The econometric analysis also seeks to explain the growth of wages and non-government employment in terms of fiscal and institutional variables. More specifically:

  • Event studies: Two event studies will be presented. In the first instance, positive labor supply shifts are defined (somewhat arbitrarily) as periods during which real productivity- and unemployment-adjusted hourly compensation rate in the business sector declined by at least 3 percent a year, for at least two consecutive years. This yields 17 periods between 1980–2003 (see Table II.4).32 The second experiment is concerned with instances of buoyant employment growth. This time, this event is defined as an increase in non-government employment by at least 2 percent a year for at least 3 consecutive years; this has the effect of isolating 11 episodes between 1980-2003 (see Table II.5).33 In each case, the behavior of a number of relevant variables—macroeconomic, fiscal, institutional, political, and social—will be examined around each particular episode: during the period, and the average of three years before, and three years afterwards.

  • Econometrics: Two simple equations are estimated, in differences. The first explores the determinants of positive labor supply shocks, defined as the change in the cyclically- and productivity-adjusted real hourly compensation, focusing on fiscal and institutional variables (Table II.6). The second equation relates the growth in non-government employment growth to wage moderation (Table II.7). In all cases, a panel model is estimated for 14 countries between 1980–2003, incorporating country fixed effects and year dummies capturing common excluded variables. Lagged dependent variables are included.34

Table II.4.

Developments Before, During, and After Wage Moderation Periods

(Event: real wages decline by at least 3 percent a year on average for at least 2 years 1/)

(Number of episodes: 17 2/)

article image
Source: Author’s calculations based on OECD, World Bank.

Defined as real productivity and unemployment-adjusted real hourly compensation rate.

Denmark, 1994-97; Denmark, 1999-00; Finland, 1993-00; Greece, 1985-86; Ireland, 1984-85; Ireland, 1987-90; Ireland, 1994-00; Italy, 1994-95; Netherlands, 1983-85; Netherlands, 1988-90; Netherlands, 1994-98; Portugal, 1985-89; Sweden, 1982-88; Sweden, 1992-95; Sweden, 1997-99; United Kingdom, 1982-88; United Kingdom, 1993-97.

Average of three years.

First year of reform period.

1 right wing; 0 centrist; -1 left wing.

Table II.5.

Developments Before, During, and After Strong Employment Growth Periods

(Event: nongovernment employment increases by at least 2 percent a year for at least 3 consecutive years)

(Number of episodes: 11 1/)

article image
Source: Author’s calculations based on OECD, World Bank.

Denmark, 1984-86; Finland, 1997-99; France 1998-00; Germany, 1989-91; Ireland, 1994-03; Netherlands, 1986-91; Netherlands, 1995-01; Spain, 1987-90; Spain, 1995-03; Sweden, 1998-01; United Kingdom, 1987-89.

Average of three years.

Adjusted for technology, hours, and unemployment.

First year of reform period.

1 right wing; 0 centrist; -1 left wing.

Table II.6.

Determinants of Wage Moderation

article image
Source: OECD
Table II.7.

Determinants of Employment Growth

article image
Source:

60. The event studies show that reform periods typically come on the heels of bad times, and are sustained through periods of buoyant growth. Following sluggish growth, both wage moderation and employment episodes generally coincide with good times, as measured by high growth rates and an improving output gap. This suggests that labor supply may indeed respond to crises, and either becomes easier to sustain during cyclical expansions, or that it itself spurs growth (which again lays the groundwork for the acceptability of further wage moderation). Moreover, there is no evidence that household savings increase during periods of wage moderation or employment growth, which could reflect the absence of adverse confidence effects; indeed, the evidence points to increasing savings beforehand, with decrease during the period itself. Finally, there is also some evidence that TFP growth is higher during the episodes in question, especially during periods of wage moderation.35 But while labor productivity growth is also higher during positive labor supply shocks, it is slower in employment booms, which is expected.

61. Consistent with the case study evidence, the event study analysis shows again that fiscal policy and labor supply improvements are intimately entwined. The overall balance clearly improves during wage moderation and employment episodes. More pertinently, falling real wages and rising employment are associated with discretionary fiscal adjustment, especially before but also during the episode in question, while this adjustment tends to peter out in the aftermath. In terms of composition, successful episodes are also associated with declining expenditure. While the earlier period tends to be one of revenue-based consolidation, expenditure declines during the labor supply event itself, including on a cyclically-adjusted basis, and including the crucial categories of transfers, social spending, and government wages. Moreover, the same key categories of expenditure all decline during the employment boom. At the same time, revenue actually falls modestly during the events, including on a cyclically-adjusted basis, and this seems solely due to taxes on labor. A number of general conclusions can be drawn. First, there is no evidence that labor market reforms are associated with fiscal policy loosening; indeed, a sound fiscal policy may be a prerequisite for successful reforms. Second, labor supply increases are more palatable alongside reduced government current expenditure and taxation.36

62. The econometric analysis also points to a clear association between labor market developments and fiscal policy (see Table II.6). What appears to matter most for wage behavior is not the stance of fiscal policy in itself, but its composition, in terms of revenue and expenditure. Discretionary increases in both expenditure and revenue appear to be associated with higher wages.37 In terms of composition, both government wages and social expenditure38 seem to be the key culprits. Likewise on the revenue side, the tax wedge (and also labor taxation directly) influences wages, but indirect and business taxation do not.39 These findings once again lend support to those (including Alesina and Ardagna, 1998, and Ardagna, 2004) who postulate a fiscal link to the labor market, insofar as cutting current expenditure (especially transfers and the government wage bill) reduces the reservation wage of union members and increases the costs of unemployment. A further conclusion to emerge is that fiscal adjustment is associated directly with employment growth, testament to the synergies that can be tapped between fiscal and labor market reforms (Table II.7).

63. Product and labor market liberalization were evident during the events in question. Product market deregulation was ongoing across these wage moderation and employment episodes, before, during, and after the events. There is no doubt that this sample of countries is heavy with product market reformers. In both event studies, there is a slight uptick in liberalization after the episode in question, although this is largely driven by liberalization bouts in the late 1980s and early 1990s in a number of countries—including Denmark, the Netherlands and the United Kingdom—that came between different wage moderation and employment expansion periods. There is also some evidence that deregulation of labor markets is associated with shifts in labor supply, mainly on the temporary side. The wage moderation period and aftermath coincide with reductions in EPLs governing temporary employment. But this is driven largely by the experiences of a handful of countries in the mid-1990s—Denmark and Sweden (during) and Italy and the Netherlands (after). In the second experiment, both permanent and temporary EPLs fall prior to the employment expansion, though this can largely be explained by developments in Spain for the former, and Spain and Sweden for the latter.

64. Evidence pertaining to changes in other labor market institutions is less clearcut. Curiously, the tax wedge appears to decline after the wage moderation period, but this is largely due to the inclusion of Greece, Italy, and Sweden in the sample, cases where wage moderation went hand-in-hand with tax wedge increases. Excluding these countries generates a decline in the tax wedge during the episodes. In the second event study, there is a clear decline in tax wedges during the employment expansion, which continues into the aftermath. The evidence pertaining to ALMPs is also unclear, with an increase before both events, and modest declines during the wage moderation and employment growth episodes themselves.40 Also, there is little relationship between changes in the replacement rate and labor supply or employment in either study. But these data do not capture crucial aspects of the benefit system, including duration and eligibility.

65. The econometric evidence also shows that only some institutional variables are significant determinants of labor supply developments. Tax wedges are important to the extent they reflect a key aspect of fiscal policy. But replacement rates, EPLs, and product market regulation do not appear to matter. The sole exception is ALMPs, which in some specifications (those that include the key expenditure and revenue variables) are associated with lower real wage growth (see Estevao, 2003, for a similar finding in a different context). Other labor market institutions that may be important, though, including benefits duration and eligibility, do not lend themselves to easy quantitative measurement.

66. Also, the feedback from labor supply shifts to employment growth depends on the degree of product and labor market regulation (Table II.7). While there is indeed a positive relationship between positive wage curve shifts and non-government employment, the interactive terms suggest that this effect is weaker in countries with heavily regulated product and labor markets (see Estevao, 2005, for a discussion of this effect as it pertains to product markets). Thus the benefits of wage moderation are greater in countries with more liberal product and labor market regulation; in this context, it may be no accident that the four case study countries are among the most liberal in Europe in this regard.

67. The events studies show that while new governments seem more inclined to implement labor supply boosting reforms, there is little evidence of other political effects. There is some evidence of a “honeymoon” effect as governments that implement policies that generate increases in labor supply tend to be newer than those before or after, defined both in terms of the tenure of the chief executive and the number of years since the last election. Breaking it down, the beginning of an episode coincided with a new chief executive (1–2 years old) in 11 of the 17 cases, and in the first year after an election in 7 cases. Newer governments are more inclined to preside over major employment expansions. There is also some evidence that wage moderation is more prevalent under right-wing parties, and that left-wing parties are more likely to be in power in the following period, although ideology does not seem to matter for employment booms. But most of the other political variables measuring government fragmentation, polarization, and instability do not appear to be relevant in either event study.41 Nor do elections seem to be any less frequent during periods of large labor supply shifts.

68. Inequality and poverty tend to nudge up in concord with labor supply in the event studies, but this is driven by a couple of countries. Extreme caution is needed in interpreting these results, given the paucity of social data. These data occur at only five year intervals, requiring interpolation for the intervening years. Clearly, then, the results are highly sensitive to both the absence of data and the choice of break point. With these caveats in mind, there is evidence of a marginal increase in inequality before and during the wage moderation period. But this is not robust, and disappears when the late 1990s Swedish episode is removed. And while the Gini coefficient also edges upwards during the employment booms, the pattern is pretty mixed across countries (with an almost equal number of increases and decreases). In the first event study, the observed uptick in poverty rates is common across all periods, with only four episodes being associated with declining poverty rates. Still, the effects are small, and driven largely by the United Kingdom experience. The same is true of the employment boom episodes, which are characterized by an increase in poverty that derives almost exclusively from the United Kingdom in the late 1980s.

H. Conclusion

69. Combining the evidence of the case studies, the event studies and the econometric analysis brings forth the following key conclusions:

  • While not necessarily comprehensive, policies tended to be internally consistent. Countries adopted a mixture of labor market, fiscal, and product market reforms that complemented and reinforced each other. But not all countries adopted the exact same strategy, especially in terms of labor market reforms; some focused on cutting benefits, others on improved targeting, on ALMPs, or lower employment protection. Cutting government expenditure and labor taxes fostered an outward shift in the labor supply curve. Low levels of product and labor market regulation also facilitated the increase in jobs rather than rents. In contrast, countries that raised labor taxes or maintained rigid product and labor markets enjoyed less success. In turn, higher employment generated further revenue which paved the way for further tax cuts and continued wage moderation—a virtuous cycle. Also, employment growth did not suppress labor productivity, which likely reflected the combination of labor and product market reforms.

  • Policies were also consistent over time. The reform process tended to be a long one in the successful countries. Once the reform engine started, there was little backtracking. Indeed, in some of the most successful cases, the same mix of fiscal and labor market policies continued unabated over two decades. In numerous countries, the size of government fell substantially over this period, and the principle of boosting labor supply through moderating wage demands became entrenched.

  • Upfront fiscal adjustment also added to the credibility of labor supply reforms. Aside from providing an environment of macroeconomic stability, fiscal adjustment created the space for further labor market reforms, by laying the groundwork for labor supply increases and facilitating the credible adoption of complementary tax cuts. Even when structural policy changed course, as in Denmark in the early 1990s, the fruits of the previous fiscal adjustment could be harvested.

  • Reform strategies that rely on coordination and competition can both be successful. The experience of Ireland and the Netherlands shows that a consensus-based approach, trading off enhanced labor supply (including through outright wage moderation) for fiscal probity and tax cuts, can be highly successful, both politically and economically. But this approach may work better in smaller, more homogeneous countries. The United Kingdom adopted an approach based on competition, curbing the power of unions and rewarding workers with tax cuts. But, as noted, inequality rose in the United Kingdom but not in those countries that relied on coordination and social partnership.

  • Successful reformers tended to neutralize opposition. Principally, this entailed giving the different parties and vested interests a stake in the process, such as by rewarding responsible wage-setting behavior by unions with labor tax cuts or by guaranteeing high benefits in the face of a strict regime and little employment protection. This meant that all kinds of government types—single-party, coalition, majority, and minority—proved capable of igniting the reform engine. Also, governments that embarked on successful reforms tended to be reelected.

  • All initial reforms started with a crisis, although they tended to continue in periods of good times. This suggests a positive feedback between the intensity of internally-consistent reforms and growth after the initial crisis. Also, successful reforms tended to involve a “fresh start” by newer governments, during their “honeymoon” period. Although the kinds of dire conditions that fueled the initial reform programs—typically two consecutive years of falling output—do not exist in Europe today, the level of unemployment does not differ much from that seen in the crisis countries at the time. This, plus the imminent onset of population aging—with its adverse implications for labor utilization, potential growth, and fiscal policy—may well provide the needed urgency for a new round of reform programs.

References

  • Alesina, Alberto, and Allan Drazen, 1991, “Why are Stabilizations Delayed?” American Economic Review, Vol. 81, pp 117088.

  • Alesina, Alberto, and Silvia Ardagna, 1998, “Tales of Fiscal Adjustment,” Economic Policy, Vol. 13, No. 27, pp 489545.

  • Alesina, Alberto, and Silvia Ardagna, Roberto Perotti, and Fabio Schiantarelli, 2002, “Fiscal Policy, Profits, and Investment,” American Economic Review, Vol. 92, No. 3, pp. 57189.

    • Search Google Scholar
    • Export Citation
  • Alesina, Alberto, Silvia Ardagna, and Francesco Trebbi, 2006, “Who Adjusts and When? On the Political Economy of Reforms,” NBER Working Paper No. 12049.

    • Search Google Scholar
    • Export Citation
  • Andersen, Torben M., 2003, “From Excess to Shortage—Recent Developments in the Danish Labour Market,” CESIFO Working Paper No. 933.

    • Search Google Scholar
    • Export Citation
  • Ardagna, Silvia, 2004, “Fiscal Stabilizations: When Do They Work and Why?” European Economic Review, Vol. 48, pp 104774.

  • Arpaia, Alfonso, and Gilles Mourre, 2005, “Labour Market Institutions and Labour Market Performance: a Survey of the Literature,” European Economy Economics Papers, No. 238, European Commission.

    • Search Google Scholar
    • Export Citation
  • Bakker, Bas B., 1999, “Crisis and Recovery” in C. Maxwell Watson, Bas B. Bakker, Jan Kees Martijn, and Ioannis Halikias [eds] The Netherlands: Transforming a Market Economy, Occasional Paper No. 81 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Bakker, Bas B., and Ioannis Halikias, 1999, “Policy Reforms and Employment Creation,” in C. Maxwell Watson, Bas B. Bakker, Jan Kees Martijn, and Ioannis Halikias [eds] The Netherlands: Transforming a Market Economy, Occasional Paper No. 81 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Barrell, Ray, and Veronique Genre, 1999, “Employment Strategies for Europe: Lessons from Denmark and the Netherlands,” National Institute Economic Review, No. 168, pp. 8298.

    • Search Google Scholar
    • Export Citation
  • Berger, Helge, and Stephan Danninger, 2005, “Labor and Product Market Deregulation: Partial, Sequential, or Simultaneous Reform?” IMF Working Paper WP/05/227 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier, 2005, “European Unemployment: the Evolution of Facts and Ideas,” NBER Working Paper No. 11750.

  • Blanchard, Olivier, and Francesco Giavazzi, 2003, “Macroeconomic Effects of Regulation and Deregulation in Goods and Labor Markets,” Quarterly Journal of Economics, Vol. 118, No. 3, pp 879907.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier, and Justin Wolfers, 2000, “The Role of Shocks and Institutions in the Rise of European Unemployment: the Aggregate Evidence,” Economic Journal, Vol. 110, No. 462, pp. C133.

    • Search Google Scholar
    • Export Citation
  • Blanchflower, David G., and Richard B. Freeman, 1993, “Did the Thatcher Reforms Change British Labour Performance,” NBER Working Paper No. 4384.

    • Search Google Scholar
    • Export Citation
  • Blank, Rebecca M., 2002, “U.S. Welfare Reform: What’s Relevant for Europe?” CESIFO Working Paper No. 753.

  • Callan, Tim, and Brian Nolan, 1999, “Income Inequality in Ireland in the 1980s and 1990s,” in Frank Barry [ed.] Understanding Ireland’s Economic Growth, London: Macmillan Press.

    • Search Google Scholar
    • Export Citation
  • Card, David, and Richard B. Freeman, 2002, “What Have Two Decades of British Economic Reform Delivered?” NBER Working Paper No. 8801.

    • Search Google Scholar
    • Export Citation
  • Castanheira, Micael, Vincenzo Galasso, Stephane Carcillo, Giuseppe Nicoletti, Enrico Perotti, and Lidia Tsyganok, 2006, “How to Gain Political Support for Reforms,” in Reforms Without Prejudice, forthcoming.

    • Search Google Scholar
    • Export Citation
  • Conway, Paul, and Giuseppe Nicoletti, 2006, “Product Market Regulation in Non-Manufacturing Sectors in OECD Countries: Measurement and Highlights,” OECD Economics Department Working Paper, forthcoming (Paris: Organization for Economic Cooperation and Development).

    • Search Google Scholar
    • Export Citation
  • Daveri, Francesco, and Guido Tabellini, 2000, “Unemployment, Growth, and Taxation in Industrial Countries,” Economic Policy, Vol. 15, No. 30, pp 47104.

    • Search Google Scholar
    • Export Citation
  • De Groot, Henri, Richard Nahuis, and Paul J. G. Tang, 2004, “Is the American Model Miss World?” CPB Discussion Paper No. 40, Netherlands Bureau of Economic Policy Analysis.

    • Search Google Scholar
    • Export Citation
  • Debrun, Xavier, and Anthony Annett, 2004, “Implementing Lisbon: Incentives and Constraints,” in Euro Area Policies: Selected Issues Paper, IMF Country Report No. 04/235 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Decressin, Jörg, Marcello Estevao, Philip Gerson, and Christoph Klingen, 2001, “Job-Rich Growth in Europe,” in France, Germany, Italy, and Spain: Selected Issues Paper, IMF Country Report No. 01/203 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Detragiache, Enrica, 2002, “Reform of the Disability Program,” in Netherlands: Selected Issues Paper, IMF Country Report No. 02/123 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Eichengreen, Barry, 1996, “Explaining Britain’s Economic Performance: A Critical Note,” Economic Journal, Vol. 106, No. 434, pp. 21318.

    • Search Google Scholar
    • Export Citation
  • Elmeskov, Jorgen, John P. Martin, and Stefano Scarpetta, 1998, “Key Lessons for Labour Market Reforms: Evidence from OECD Countries’ Experiences,” Swedish Economic Policy Review, vol. 5, pp. 20552.

    • Search Google Scholar
    • Export Citation
  • Estevao, Marcello, 2003, “Do Active Labor Market Policies Increase Employment,” IMF Working Paper No. 03/234 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Estevao, Marcello, 2005, “Product Market Regulation and the Benefits of Wage Moderation,” IMF Working Paper No. 05/191 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • European Commission, 2006, “Public Finances in EMU, 2003,” European Economy, No. 3.

  • Fernandez, Raquel, and Dani Rodrik, 1991, “Resistance to Reform: Status Quo Bias in the Presence of Individual-Specific Uncertainty,” American Economic Review, Vol. 81, No. 5, pp 14655.

    • Search Google Scholar
    • Export Citation
  • Gaard, Soren, and Mads Kieler, 2004, “Two Decades of Structural Reform in Denmark: a Review,” unpublished, Ministry of Finance, Denmark.

    • Search Google Scholar
    • Export Citation
  • Garibaldi, Pietro, and Paolo Mauro, 2002, “Anatomy of Employment Growth,” Economic Policy, No. 34, pp 69113.

  • Genre, Veronique, Ramon Gomez Salvador, and Ana Lamo, 2005, “European Women: Why Do(n’t) They Work?” ECB Working Paper, No. 454.

  • Grubb, David, 2000, “Eligibility Criteria for Unemployment Benefits,” OECD Economic Studies No. 31 (Paris: Organization for Economic Cooperation and Development).

    • Search Google Scholar
    • Export Citation
  • Hallerberg, Mark 2004, Domestic Budgets in a United Europe: Fiscal Governance from the End of Bretton Woods to EMU, Cornell University Press.

    • Search Google Scholar
    • Export Citation
  • Haque, Nadeem Ul., M. Hashem Pesaran, and Sunil Sharma, 1999, “Neglected Heterogeneity and Dynamics in Cross-Country Savings Regressions,” IMF Working Paper No. 99/128 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Hartog, Joop, 1999, “Whither Dutch Corporatism?, Or: A Turbulent Tango for Market and State,” Institute for Research on Poverty Discussion Paper No. 1197–99, University of Amsterdam.

    • Search Google Scholar
    • Export Citation
  • Hauptmeier, Sebastian, Martin Heipertz, and Ludger Schuknecht, 2006, “Expenditure Reform in Industrialized Countries: A Case Study Approach,” ECB Working Paper, No. 634.

    • Search Google Scholar
    • Export Citation
  • Honohan, Patrick, and Brendan Walsh, 2002, “Catching Up With the Leaders: the Irish Hare,” Brookings Papers on Economic Activity, No. 1, pp 177.

    • Search Google Scholar
    • Export Citation
  • Hunt, Ben, 2004, “The Role of Social Partnership Agreements in Ireland: Contributing to the Boom and Facilitating Adjustment to Sustainable Growth,” in Ireland: Selected Issues Paper, IMF Country Report No. 04/349 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • International Monetary Fund, 2004, “Fostering Structural Reforms in Industrial Countries,” in World Economic Outlook, April (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Jaumotte, Florence, 2003, “Female Labour Force Participation: Past Trends and Main Determinants in OECD Countries,” OECD Economics Department Working Paper No. 376.

    • Search Google Scholar
    • Export Citation
  • Judson, Ruth A., and Ann L. Owen, 1999, “Estimating Dynamic Panel Data Models: A Guide for Macroeconomists,” Economics Letters, Vol. 65, pp. 915.

    • Search Google Scholar
    • Export Citation
  • Maher, Maria, and Michael Wise, 2005, “Product Market Competition and Economic Performance in the United Kingdom,” OECD Economics Department Working Paper No. 433 (Paris: Organization of Economic Cooperation and Development).

    • Search Google Scholar
    • Export Citation
  • Martin, John P., and David Grubb, 2001, “What Works and For Whom: A Review of OECD Countries’ Experiences with Active Labour Market Policies,” Swedish Economic Policy Review, Vol. 8, pp. 956.

    • Search Google Scholar
    • Export Citation
  • Nickell, Stephen, and Glenda Quintini, 2002, “The Recent Performance of the UK Labour Market,” Oxford Review of Economic Policy, Vol. 18, No. 2, pp. 20220.

    • Search Google Scholar
    • Export Citation
  • Nickell, Stephen, and Jan van Ours, 2000, “The Netherlands and the United Kingdom: A European Unemployment Miracle?” Economic Policy, Vol.15, No.30, pp. 13580.

    • Search Google Scholar
    • Export Citation
  • Nickell, Stephen, Luca Nunziata, and Wolfgang Ochel, 2005, “Unemployment in the OECD Since the 1960s: What Do We Know?” Economic Journal, Vol. 115, pp. 127.

    • Search Google Scholar
    • Export Citation
  • Nicoletti, Giuseppe, and Stefano Scarpetta, 2005, “Product Market Reforms and Employment in OECD Countries,” OECD Economics Department Working Paper No. 472 (Paris: Organization for Economic Cooperation and Development).

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Cooperation and Development, 2004, “Employment Protection Regulation and Labour Market Performance,” in Employment Outlook, Paris: OECD.

    • Search Google Scholar
    • Export Citation
  • Organization for Economic Cooperation and Development, 2006, “Political Economy of Structural Reform,” Working Paper No. 1 on Macroeconomic and Structural Policy Analysis, Paris: OECD.

    • Search Google Scholar
    • Export Citation
  • Pissarides, Christopher, 2003, “Unemployment in Britain: A European Success Story,” CESIFO Working Paper No. 981.

  • Sapir, Andre, 2005, “Globalisation and the Reform of the European Social Models,” unpublished, Bruegel Institute, Brussels.

  • Swagel, Philip, 2000, “Welfare and Labor Market Reform in the United Kingdom: Evidence from the International Experience,” in United Kingdom: Selected Issues Paper, IMF Country Report No. 00/106 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Tanzi, Vito, and Ludger Schuknecht, 2005, “Reforming Public Expenditure in Industrialized Countries—Are There Trade-Offs?” ECB Working Paper, No. 435.

    • Search Google Scholar
    • Export Citation
  • Westergaard-Nielsen, Niels, 2001, “Danish Labour Market Policy: Is It Worth It?” Centre for Labour Market and Social Research Working Paper No. 01–10.

    • Search Google Scholar
    • Export Citation

Appendix I. Reforming Labor Market Institutions in the Case Study Countries

70. A large number of studies find that a bevy of labor market institutions may have been responsible for poor labor market outcomes in Europe. The institutions under indictment typically include generous benefits with long duration, high tax wedges, restrictive employment protection legislation (EPL), and union power (see Arpaia and Mourre, 2005, for a comprehensive overview of the state of play in this area). The negative supply shocks of the late 1970s- early 1980s may have interacted with these prevailing institutions to lead to persistently high unemployment (Blanchard and Wolfers, 2000). Or governments may have responded to adverse shocks by using institutions to protect workers, a trend that was only partially reversed after the 1980s (Blanchard, 2005). In a fairly representative result, Nickell, Nunziata, and Ochel (2005) argue that institutions can explain slightly more half of the increase in European unemployment between the 1960s and 1990s, with a division between unemployment benefits (39 percent), labor taxes (26 percent), union power (19 percent), and EPLs (16 percent).

Tax wedges

71. High tax wedges have been frequently singled out as a leading cause of poor labor market outcomes in Europe. Nickell, Nunziata, and Ochel (2005) claim that they rank second in importance behind high unemployment benefits. Garibaldi and Mauro (2002) argue that employment growth is associated primarily with low taxes and low dismissal costs. And Daveri and Tabellini (2002) go even further, assigning the brunt of the blame for high unemployment squarely on labor taxes, especially in intermediate cases of wage bargaining.42

72. All countries embarked on labor tax reforms that had the effect of reducing tax wedges.

  • In Ireland, taxes on labor declined steadily over the 1990s, as part of the centralized social partnership agreements; the top income tax fell from 68½ percent in the mid-1980s to 42 percent by the end of the century. As well as declining marginal rates, income tax thresholds were raised in real terms, removing many of the low paid from the tax net altogether.

  • The Netherlands took a similar route. Social security contributions were cut dramatically (by 58 percent) for low-income workers, and if these workers came from long-term unemployment, they were exempted from contributions during the first four years of employment.

  • In the United Kingdom, the top marginal rate fell from 83 percent in 1975 to 60 percent in 1985 and 40 percent by 1990. The tax system shifted from a household to an individual basis, and tax brackets were lowered as the government moved toward indirect taxation.

  • Although it was not key to its strategy, Denmark undertook some tax reform in the 1990s to broaden the base and reduce rates, even though its taxation ratio remains among the highest in Europe. While lower than in the early 1990s, marginal tax rates remain high, exceeding 60 percent for high earners and 50 percent on average (Gaard and Kieler, 2004).

Benefit reform

73. Benefits reform had a tripartite character, focusing on reducing generosity, shortening duration and tightening eligibility requirements. Reforms were especially prominent in Denmark, the Netherlands, and the United Kingdom. Benefit reform was not a major aspect of the Irish story.

74. Benefits generosity was lowered in the Netherland and the United Kingdom. In the Netherlands, the replacement rate for unemployment, sickness, and disability benefits was reduced from 80 percent to 70 percent of the last earned wage in 1986, and the disabled unemployed no longer received full benefits. The generosity of unemployment benefits in the United Kingdom was also reduced, owing to the abolition of the earnings related supplement (the only component tied explicitly to past earnings), the suspension of statutory indexation, the elimination of child dependent allowances, the taxing of unemployment benefits, and the removal of any entitlement for under-18s.43 Denmark stands out as the country with a continuing attachment to a generous benefits regime. Benefits levels were kept at 90 percent of previous earnings subject to a maximum benefit level (meaning that replacement rates are highest for low earners and decline with income) and not reduced over time. This gives rise to clear disincentive effects.44

75. The maximum duration of benefits was lowered in Denmark and the Netherlands. This was a key aspect of the Danish reforms. Whereas the maximum benefit period had been virtually open ended, it was set to seven years in 1994, reduced to five years in 1998, and again to four years in 2000. Duration was also cut in the Netherlands, from 2½ years to six months for workers under 23, and less severely for other age groups (except for those over 58 who received six years of eligibility). Neither Ireland nor the United Kingdom engaged in this aspect of reform.

76. Reforms geared toward tightening eligibility conditions were widespread.

  • In Denmark, the work requirement to qualify for unemployment benefits was raised from six months to one year in the preceding three years, and the loophole whereby participation in ALMPs could be used to extend the benefit period was eliminated. Eligibility could now only be restored with a regular unsubsidized job.

  • In the Netherlands, eligibility also became progressively tighter. By 1995, the unemployed needed to stand ready to accept a broader range of jobs or enter training, and were required to accept any job offer after 18 months of unemployment. By 1998, the sanction rate—defined as sanctions during benefit periods as a percentage of the average stock of benefit claims—was 36 percent, the highest in the OECD (Nickell and van Ours, 1999). Also, applicants for disability were gauged on whether they could perform any job, regardless of training or employment history, and those under 50 were reexamined.

  • In the United Kingdom, a Restart program was introduced in 1986, which called for compulsory counseling and referral for those in unemployment for more than six months. From 1988, an interview was required every six months. In 1989, the unemployed needed to show good cause before refusing work. In 1991, participation in a labor market program was made compulsory for those in unemployment for more than two years. Under the “Stricter Benefit Regime” the number of sanctions doubled in 1994 from a few years earlier (Grubb, 2000).

  • Irish efforts in this area came limited and late. From 1998, criteria were toughened, as all those unemployed for longer than six months were required to attend interviews in an attempt to improve matching or identify training needs.

77. The conditioning of unemployment benefits on activation measures was a key component of reforms in Denmark in particular. In a nutshell, the unemployed were required take part in ALMPs. This requirement kicked in after three years in unemployment in 1996, and was reduced gradually until it became one year in 2002, so that the activation period constituted 75 percent of the benefit duration period. Youth unemployment was particularly targeted in that beneficiaries aged 25 or less without qualifying education could no longer receive passive benefits longer than 26 weeks. Beyond that period, they were obliged to either work or further their education at a benefit level around half of the unemployment benefit. Non-compliance in the youth schemes brought benefits to an end. The United Kingdom adopted a similar experiment in 1998, as participation in ALMPs was required for youth in unemployment status after 6 months.

Active labor market policies

78. Numerous studies have found a positive effect of Active Labor Market Programs (ALMPs) on labor market outcomes. Various theoretical reasons suggest that expenditure on ALMPs could affect employment and wages by: (i) shifting out the labor demand curve, by boosting labor productivity, and by achieving better matching between vacancies and the unemployed; (ii) shifting down the wage or labor supply curve by better matching and keeping unemployed workers attached to the labor force; (iii) shifting up the wage or labor supply curve by lowering the disutility of unemployment and raising the reservation wage (see Estevao, 2003). In a comprehensive analysis, Estevao (2003) found that ALMPs raised employment, especially in the case of subsidized employment, while job search assistance and youth measures were actually counterproductive. Also, ALMPs fostered wage moderation. Evidence from the United States also supports the view that work experience trumps training (Blank, 2002). Martin and Grubb (2001) believe that job search assistance and wage subsidies can be effective, but the impacts tend to be small. Going even further, De Groot, Nahuis, and Tang (2004) argue that the use of ALMPs permits countries to enjoy both high employment and high equity.

79. Countries differ widely in size and composition of ALMPs. While the average EU country spends roughly 1 percent of GDP on active measure, the Netherlands and Denmark lead the pack with around 1¾ percent and 1 ½ percent respectively, followed by Ireland at 1¼ percent, with the United Kingdom near the bottom (½ percent). With its long history of apprenticeship, job training has been the chief activation measure in Denmark, although the education subsidy gained prominence in the 1990s, prompted by the more rigorous youth activation requirements. The focus on the Netherlands is mostly on measures for the disabled, but job training initiatives have grown over time. Ireland’s main interest has been on subsidized employment. The United Kingdom concentrates far more limited resources on administrative services and youth measures. In 1997, it introduced its “New Deal for Young People”, which established mandatory activation for receipt of unemployment benefits. After a 4-month gateway, the young unemployed needed to choose between education or training, subsidized employment, working for a non-profit entity, or participating in an environmental task force.

Employment protection legislation

80. The evidence pertaining to the effect of employment protection on labor market performance is decidedly mixed. By decreasing both hiring and layoffs, the effect of EPLs on the unemployment rate could be ambiguous. But while Nickell, Nunziata, and Ochel (2005) assign only a minor role to EPLs in explaining the rise in European unemployment over time, others adopt a different perspective (Garibaldi and Mauro, 2002, for example). Elmeskov, Martin, and Scarpetta (1998) believe that EPLs are most detrimental in the intermediate bargaining case. In a comprehensive evaluation, OECD (2004) finds that EPLs reduce the employment rate among youth and women, but not prime-age men—a classic insider-outsider scenario.

81. The case study countries stand out in terms of low EPLs, and both Denmark and the Netherlands undertook reforms. Since the late 1980s (when data became available), EPLs were lowered in both Denmark in the Netherlands, pushing them toward the bottom, while Ireland and the United Kingdom changed little, yet maintaining their relative positions. The Danish liberalization in the 1990s reflects the expanding role of temporary work agencies (OECD, 2004). Dutch reform began in the late 1980s, and focused on streamlining laying off rules and relaxing hiring procedures (see Barrell and Genre, 1999). Working-time rules became more flexible, which had the effect of encouraging part-time work. Overtime legislation was also relaxed. Further reforms of fixed-terms contracts, probation periods, and dismissal procedures took place in the late 1990s. In particular, temporary work agencies gained more prominence as regulations relating to them were relaxed; this contributed to the explosion of part-time work and flexible contracts, especially among women.

17

Prepared by Anthony Annett.

18

The literature shows that female participation is affected positively by such factors as maternity leave, childcare costs, and availability of part-time work—all part of the Nordic social model (see Jaumotte, 2003; Genre, Salvador, and Lamo, 2005). Denmark remains the leader of the pack in this regard, while both Ireland and the Netherlands lagged.

19

Following a “baby bulge” in the 1960s and 1970s, the birth rate declined toward average European levels by 1980. Emigration, especially to the United Kingdom, was also a constant in Ireland in the postwar period, giving rise to a large pool of potential returning immigrants.

20

The results do not change if business employment is used instead.

21

The number of welfare recipients—enrolled in unemployment, sickness, and disability programs—more than doubled, from 8 percent of working age population in 1970 to 18 percent by 1982 (Bakker, 1999).

23

The hourly compensation rate in the business sector is deflated by the private consumption deflator and an index of labor-augmenting technical progress. But since changes in wages may simply reflect cyclical changes in unemployment, or a movement along the wage (labor supply) curve, the growth in wages is also adjusted for the change in the unemployment rate, based on an elasticity of wage costs to unemployment of 0.1, a standard assumption in the literature.

24

Membership in unions declined from around 48 percent of the labor force in 1980 to around 40 percent a decade later (Honohan and Walsh, 2002).

25

Early industrialization provided the United Kingdom with a particular set of institutions that were never reformed as it was undefeated in the war: a banking system specializing in trade credits instead of industrial finance; an industrial structure dominated by single-plant firms; and a fragmented craft-based unionism (Eichengreen, 1996).

26

The agreements are: (i) the Programme for National Recovery, 1987–91; (ii) the Programme for Economic and Social Partnership, 1991-93; (iii) the Programme for Competitiveness and Work, 1994–96; (iv) Partnership 2000, 1997–99; (v) the Programme for Prosperity and Fairness, 2000–02; (vi) Sustaining Progress, 2003–05. For details, see Hunt (2004).

27

See Appendix I for more in-depth discussion of the different reform programs.

28

The tax wedge is defined as employees’ and employers’ social security contributions and personal income tax less transfer payments as percent of gross labor costs, averaged between single and married workers.

29

This replacement rate is defined as the average over three different family types (single, married with dependent spouse, and married with working spouse), and two different income levels (67 percent of average earnings, and 100 percent of average earnings). The pattern does not change much by aggregating further and taking three different time periods—the first year, the second and third years, and the fourth and fifth years.

30

Ireland’s reforms concentrated on administrative simplification and openness to trade and foreign direct investment (Nicoletti and Scarpetta, 2005).

31

In their analysis, ambitious reformers included Ireland, the Netherlands, and the United Kingdom in the 1980s, joined by Belgium, Finland, Spain, and Sweden in the 1990s.

32

Consideration was also given to other definitions, including a more liberal one (an average of 2 percent a year), which threw up 25 episodes. The results were unchanged.

33

A less strict version of the analysis was also conducted (1 percent a year), which doubled the number of episodes, but altered none of the qualitative conclusions. Nor did the results change by looking at the nongovernment employment rate.

34

Judson and Owen (1999) argue that when the time series is long enough relative to the cross-section dimension, the bias inherent in dynamic panel estimation is not large enough to make alternative estimators more desirable. Others have also argued that when the time span covered by the data is reasonably large (around 22), then the application of IV-type estimators to a first differenced version of the dynamic panel model does not seem necessary, and can even lead to a large loss of efficiency (see Haque, Pesaran, and Sharma, 1999).

35

This is partly mechanical for the wage moderation episodes, however, as real wages are adjusted for TFP growth.

36

An event study undertaken by Hauptmeier, Heipertz, and Schuknecht (2006)—focusing instead on incidents of large fiscal adjustment—arrived at similar conclusions: cutting transfers and public consumption is typically part of a comprehensive package that includes labor market reforms and is associated with higher growth, employment, and consumer confidence.

37

The same is true for unadjusted revenue and expenditure, in unreported results.

38

Public social expenditure includes spending on pensions, unemployment benefits, sickness and disability programs, and healthcare.

39

This result is also not reported in Table II.6.

40

It should be noted that data on EPLs and ALMPs are only available from 1985, curtailing the sample.

41

The irrelevant variables are: (i) effective number of parties in government and legislature, defined as the inverse of the sum of squared seat shares held by each party; (ii) government and legislative fractionalization, which is the probability that two representatives picked at random will be from different parties; (iii) the size of the government majority; (iv) polarization, defined as the maximum ideological difference between the executive party and the four principle parties in the legislature; (v) stability, which measures the percent of veto players who drop from a government in any given year. The source for all political variables is the World Bank’s Database of Political Institutions.

42

Neither decentralized, where labor taxes are borne by workers, nor centralized where unions take employment effects into account when bargaining over wages.

43

For details, See Blanchflower and Freeman (1993) and Nickell and Quintini (2002).

44

Estimates suggest that 23 percent of employed women and 12 percent of employed men earn little extra from working as opposed to receiving benefits (Westergaard-Nielson, 2001).