This Selected Issues paper on Sweden examines developments in the government structural balance. The paper identifies the respective contributions of cyclical and structural factors in the present fiscal consolidation phase. It reviews recent developments and the medium-term fiscal outlook. The paper presents estimates of the structural and cyclical components of the government financial balance. The factors responsible for the fiscal improvement anticipated over the medium term are analyzed. The paper also analyzes the medium-term outlook for public debt, with particular attention to the sensitivity of debt dynamics to different macroeconomic conditions.

Abstract

This Selected Issues paper on Sweden examines developments in the government structural balance. The paper identifies the respective contributions of cyclical and structural factors in the present fiscal consolidation phase. It reviews recent developments and the medium-term fiscal outlook. The paper presents estimates of the structural and cyclical components of the government financial balance. The factors responsible for the fiscal improvement anticipated over the medium term are analyzed. The paper also analyzes the medium-term outlook for public debt, with particular attention to the sensitivity of debt dynamics to different macroeconomic conditions.

VI. The Swedish Labor Market: The Challenge of Reform29

A. Introduction

113. The performance of the Swedish labor market deteriorated sharply in the early 1990s. After having been very low for decades, open unemployment rose more than fivefold between 1991 and 1993 to reach continental European levels (Chart 1). The deep economic recession of the period clearly played a role in the rise in unemployment. Yet, favorable growth since 1993 has failed to make a dent in the high unemployment. With fiscal consolidation on a firm track and inflation under control, lowering unemployment has become the main priority of policy in Sweden. This paper explores how this challenge can be met.

Chart 1.
Chart 1.

Sweden: Labor Market Developments

Citation: IMF Staff Country Reports 1997, 096; 10.5089/9781451835847.002.A006

Sources: Statistics Sweden; and OECD.1/ Open unemployment plus labor market measures.

114. The first section below describes developments in the Swedish labor market in the 1990s and identifies the main characteristics of the unemployment problem. The next section gives an account of the factors that were chiefly responsible for the rise in unemployment and seeks an explanation of the persistence of high unemployment in recent years. The third section underscores that Sweden’s high unemployment, while of relatively recent vintage, is closely related to the unemployment problem many continental European countries have faced over the last two decades, and that both require similar responses for a lasting solution. The last section reviews efforts to reform labor markets in the Netherlands over the last decade and draws some lessons for Sweden.

B. Labor Market Developments in the 1990s

115. From the late 1970s through the 1980s, the Swedish labor market offered contrasting developments to those in most European countries. While unemployment in the latter rose inexorably to the double digit level, unemployment in Sweden remained fairly stable between 2–3 percent. The vaunted “Swedish Model,” which combined active labor market policies, an unemployment insurance system with a fixed duration of benefits, and centralized wage bargaining to deliver wage moderation, appeared to be performing well. However, when the 1990s came around, this would no longer be the case.

116. In 1990, open unemployment in Sweden was 1.5 percent and a further 3 percent of the labor force were engaged in various labor market programs for a total unemployment rate of 4.5 percent;30 labor force participation stood at 85 percent. By 1993, open unemployment had risen to 8.2 percent and participation in labor market programs to 5.1 percent for total unemployment of 13.3 percent; labor force participation had fallen to 79 percent. All in all, employment contracted by 12 percent between 1990 and 1993, while GDP contracted by less than 5 percent.31 The contraction in employment was spread across all sectors, but was most pronounced in manufacturing and construction which saw 25—30 percent of their jobs disappear, accounting for around two thirds of the decline in employment. Public sector employment also declined, accounting for around 20 percent of the total contraction.

117. For the period 1993–96, employment has been stagnant despite cumulative GDP growth of around 8 percent, giving rise to the phenomenon of “jobless growth.” The modest employment gains in the booming manufacturing export sector, which was given a boost by the sharp real depreciation of the krona in 1993 following the de-linking from the ECU in late 1992, have been offset by continued employment declines in the public sector, especially local governments, associated with the ongoing fiscal consolidation.32 The overall result has been that open unemployment has been fairly stable at around 8 percent and total unemployment at close to 13 percent.

118. Unemployment in Sweden shares many of the characteristics that are well known features of unemployment in other European countries. Unemployment is highest among young and unskilled workers and the incidence of long-term unemployment has been on the rise. Open employment among 16–24 year olds is about twice the average. Similarly, unemployment among those with only primary and secondary education is around twice that of those who have completed university education. Furthermore, the incidence of long-term unemployment33 in open unemployment has risen from around 5 percent in the late 1980s to 16 percent in 1995; given that participation in a labor market program qualifies an unemployed person for a new spell of unemployment benefits, the real incidence of long-term unemployment is most likely significantly understated.

119. During the 1970s and 1980s, inflation in Sweden was relatively high but did not exhibit an accelerating trend. Furthermore, inflation tended to fall during periods of higher unemployment and pick up during periods of lower unemployment. This suggests that the NAIRU was not far from the average rate of open unemployment of 2–3 percent in this period. However, various estimates, by the staff and others (for example, OECD, 1996), indicate that the NAIRU may now have risen to 6–8 percent. Admittedly, rising estimates of the NAIRU tend to track the actual unemployment rate and may, therefore, not represent estimates of some fundamental rate of equilibrium unemployment. They, nevertheless, point to the potential persistence of a particular level of unemployment over a long period.

C. Accounting for the Rise and Persistence of Unemployment34

120. The most plausible story behind the sudden rise in unemployment in Sweden in the early 1990s involves a host of cyclical and structural factors that came together at a particular point in time. Some of the latter have subsequently contributed to the persistence of the high level of unemployment and jobless growth.

121. At the end of the 1980s, the Swedish economy was seriously overheated as a result of a credit boom related to financial deregulation. Household saving was negative by 5 percent of disposable income and housing prices were at a record high. The major tax reform of 1990–91, which lowered marginal income tax rates but also raised after-tax real interest rates, had a role in bursting this bubble. The resultant fall in housing prices along with a rise in precautionary saving in the face of rising unemployment and emerging fiscal deficits led to a rise in the saving rate to 8 percent of disposable income by 1992, and a collapse in domestic demand with a loss of jobs in sectors oriented towards the domestic market. To some extent a vicious circle was at work: rising unemployment led to increased precautionary saving which lowered domestic demand which in turn led to greater job losses. In recent years, private consumption demand has remained subdued owing to the effects of the strong fiscal consolidation program on household disposable income through higher taxes and lower transfers.

122. Also in the late 1980s and early 1990s, the manufacturing export sector suffered from a lack of competitiveness because of high domestic cost levels and weakening international demand. Furthermore, in 1991, the authorities signaled a renewed commitment to low inflation when they linked the krona to the ECU. This represented a break with the accommodative policies of the past, under which cost competitiveness was maintained through periodic devaluations. Under these conditions, and faced also with higher real interest rates, firms were forced to seek greater labor productivity in order to stay competitive and profitable. Labor shedding by firms was given a further boost when they were made responsible for the first two weeks of sickness benefits and thus given a strong incentive to tackle a rampant problem of absenteeism. Increased concerns about job security also made workers more willing to work longer hours, further depressing the demand for new hires. Thus, the boom in manufacturing exports in 1994–95, resulting from the real depreciation of the krona in 1993, reversed less than one third of the job losses in manufacturing experienced earlier; manufacturing employment then declined again marginally in 1996.

123. It is important to note that all the employment generation in Sweden in the 1970s and 1980s took place in the public sector, making public sector employment at close to 40 percent of the total the highest among the industrial countries. The rapid emergence of fiscal deficits in the recession of the early 1990s and the explosion of public debt made it clear that the limits of public sector expansion had been surpassed. Since the early 1990s, with fiscal consolidation under way, public sector employment has contracted by more than 10 percent, first in central administration but more recently in local governments. The continued decline in public sector employment has contributed to the persistence of high unemployment.

124. Fully centralized wage bargaining in Sweden broke down in the early 1980s, with wage setting since taking place at the industry and local level. This intermediate form of wage bargaining would a priori be expected to lead to a poor trade-off between wage increases and employment creation because of the familiar insider-outsider problem, in which employed persons do not incorporate fully in their real wage demands the costs to the unemployed. The insider-outsider problem should be magnified by arrangements such as those in Sweden under which government bears almost the entire cost of unemployment insurance. There is scant evidence of the insider-outsider problem in Sweden in the 1980s, when unemployment rates were low, although wage increases were at times excessive leading to cost pressures and an erosion of competitiveness. More recent wage developments may provide evidence of the existence of the problem. In 1996, nominal wage increases in manufacturing exceeded 6 percent, despite the very high levels of unemployment. The evidence is not clear cut, however, in that expectations of rising inflation may have played a role in the excessive wage demands; with the recent fall in inflation expectations wage demands have come down. Nevertheless, the high wage increases of 1996 may have contributed to the recent decline in manufacturing employment and the persistence of high unemployment.

125. It is difficult to argue that Sweden’s generous system of unemployment insurance, which at the beginning of the 1990s replaced 90 percent of the earlier income of the unemployed up to a level corresponding to the average manufacturing wage, directly contributed to the rise in unemployment because it had been around for decades, well before the sudden emergence of high unemployment. However, the system was not designed to handle high levels of prolonged unemployment and may have been overwhelmed in the recent period. When unemployment was low, the public employment services were able to strictly enforce the requirement that an unemployed person, with some limitations, accept an available job; this may no longer be the case. The unemployment insurance system—coupled with active labor market measures, participation in which re-qualifies for unemployment benefits—may have become a refuge for some who, especially with the stigma of unemployment receding, genuinely do not seek work and may thus be contributing to unemployment persistence. This may be the case despite the lowering of the replacement rate from 90 percent to 75 percent under the fiscal consolidation program.35

126. It is also difficult to attribute the rise in unemployment to Sweden’s relatively restrictive job protection regime based on the last-in-first-out principle. However, the job protection provisions may pose a serious obstacle to employment creation, especially for small- and medium-sized enterprises, which in some countries, such as the United States, have been the locomotive behind employment growth.

127. Sweden certainly has some of the largest tax wedges between labor costs and take-home pay among the industrial countries. However, at the time unemployment exploded in the early 1990s these wedges were actually being narrowed as part of the 1990–91 tax reform. More recently, the tax wedges have widened again as a result of tax increases in the fiscal consolidation program, which may have pushed take-home pay below the reservation wage for some workers, pushing them into unemployment.

128. There is some evidence that the relationship between unemployment and vacancies (the Beveridge curve) has deteriorated in recent years, pointing to a rising mismatch between labor supply and labor demand. The average skill level required to fill announced vacancies has also clearly risen. This may point to an increasing lack of highly skilled workers, a problem that may be related to the solidaristic wage compression, for which there is a long standing tradition in Sweden, and the resulting low return on investment in education and human capital. It may also account for the rising unemployment among the young and unskilled.

129. In sum, the sudden rise of unemployment in Sweden appears to have been the result of a complex interaction of cyclical and structural factors that came together at a particular point in time. Although important institutional arrangements and legal provisions for the labor market—such as concerning wage formation, unemployment insurance, job protection and taxes on labor—may not explain the timing and size of the rise in unemployment in the early 1990s, they are likely to have contributed to the persistence of high unemployment in the recent period despite adequate growth.

D. Swedish Unemployment in a European Perspective

130. For most of the 1970s and 1980s, the Swedish labor market appeared to perform much better than the labor markets of most European countries, and was even held up as an example for other countries to emulate. This was the case despite the fact that Sweden’s labor market harbored to a great extent the same features which more recently have been blamed for the poor performance of European labor markets, including incompletely centralized or decentralized wage bargaining, overly generous unemployment benefits in terms of both replacement ratios and duration, excessively stringent job protection provisions and overly high tax wedges on labor (OECD 1994, OECD 1997). In retrospect, it can be seen that Sweden was able to preserve low employment by pursuing more accommodative policies than other European countries and by tolerating a greater expansion of the public sector, and the unavoidable rise in the level of taxation. As a result, Sweden did not experience the ratcheting up of unemployment seen in other European countries in the wake of the two oil-related supply shocks of the 1970s and 1980s and the attempt to rein in inflation in the 1980s. Sweden did not commit to low inflation until the early 1990s when an overheated economy and a bloated public sector quickly led to a deep recession, a fiscal crisis and an explosion of unemployment. The persistence of high unemployment in Sweden in the non-accommodative policy environment of the recent period may to a considerable extent be blamed on those features of the Swedish labor market it shares with other European labor markets.

131. Sweden’s unemployment problems are thus closely related to those of other European countries and will require a similar response for a lasting solution. In a fundamental sense, labor markets need to function to allow labor to gradually be reallocated in response to sector-specific shifts in technology, changes in the composition of demand and new trends in international competition (Lindbeck, 1997). In effect, labor markets need to be flexible. Among the important requirements is that education, training and not least relative wages adjust to changes in relative demand to avoid increased mismatches between demand and supply for workers with different skills.

132. A number of the characteristics of European labor markets, including the Swedish one, militate against labor market flexibility, including the presence of the insider-outsider problem, which prevent the outsiders from pricing themselves into employment and an over-emphasis on wage compression, which reduces incentives for investment in human capital by limiting the return on education. Furthermore, overly generous unemployment benefits discourage job search while job protection provisions raise hiring costs and discourage job creation, and wide tax wedges between labor costs and take-home pay discourage both labor supply and labor demand. The need for more flexible labor markets is amplified by the trend for job creation in the advanced economies which is likely to be primarily in the service sector (IMF, 1997).

E. Lessons from the Netherlands

133. The causes of stagnant employment and persistent unemployment in Sweden thus differ little from the causes of similar problems in other European countries. Unfortunately, the European experience offers few examples of successful labor market reforms; the Netherlands may be the exception.

134. The Netherlands encountered their labor market crisis in the early 1980s—a decade earlier than Sweden—also in the context of a fiscal crisis, which contributed to the formation of a consensus on the need for wage moderation, fiscal consolidation, social insurance and labor market reforms. The labor market reforms focused on measures to stimulate both the supply of labor and job search—including lowering of the replacement ratio for unemployment benefits and shortening the benefit period, tightening eligibility criteria and cutting the income tax burden—and labor demand—including cutting the real minimum wage and lowering payroll taxes and other non-wage labor costs. Wage moderation was triggered by a change in labor union attitudes, but fiscal consolidation and labor market reforms played a role in sustaining the change in wage behavior.

135. The reform efforts in the Netherlands over the past decade have led to a steady narrowing of the labor tax wedge and relatively moderate real wage developments (Chart 2). The benefits have also been clear in terms of a strong rise in employment and an easing of the unemployment rate. These developments contrast sharply with corresponding developments in Sweden, where the already very wide labor tax wedges have again widened in recent years as a result of income and payroll tax increases under the fiscal consolidation program and real wage growth has accelerated notwithstanding the very high levels of unemployment, reflecting the strong profitability of manufacturing firms and the potential presence of insider-outsider problems.

Chart 2.
Chart 2.

Sweden: Labor Market Comparisons

Citation: IMF Staff Country Reports 1997, 096; 10.5089/9781451835847.002.A006

Sources: IMF, World Economic Outlook; OECD, Tax/Benefit Positions of Production Workers (various issues).1/ Difference between employment andoutput growth.2/ Open unemployment plus labor market measures.3/ Income taxes and employer and employee social security contributions in percent of gross labor costs for overage production workers (one-earner couple with two children).4/ Hourly wages deflated by the CPI.

136. The labor market reforms in the Netherlands are no panacea and labor market conditions there are far from perfect; for example, the labor force participation rate, especially for women, remains quite low, and well below that in Sweden. Nevertheless, the experience of the Netherlands shows that comprehensive and consistent reforms aimed at making labor markets more flexible will over time produce beneficial results in terms of rising employment and declining unemployment. The relevance for Sweden of the labor market reforms in the Netherlands should be enhanced by the consensus orientation of both societies and their concern with egalitarian outcomes.

STATISTICAL APPENDIX

Table A1.

Sweden: GDP and Expenditure Components

article image
Sources: Statistics Sweden; and IMF, World Economic Outlook.

Changes in percent of previous year’s GDP.

Weighted by the geographic distribution of Sweden’s 1986–88 average exports.

Table A2.

Sweden: Household Disposable Income, Consumption, and Saving

article image
Sources: National Institute of Economic Research; and data provided by the Swedish authorities.
Table A3.

Sweden: Gross Fixed Investment

article image
Source: Statistics Sweden.
Table A4.

Sweden: Output, Employment (Hours), and Productivity by Sector 1/

article image
Source: National Institute of Economic Reserach.

Output estimated from the production side in value added at 1991 producer prices. Productivity measured in hours worked.

Table A5.

Sweden: Labor Market

article image
Sources: Statistics Sweden; Ministry of Finance, Sweden’s Economy, and data provided by the Swedish authorities.
Table A6.

Sweden: Average Hourly Earnings, Wage Costs, and Unit Labor Costs

(Annual percentage changes)

article image
Sources: National Institute of Economic Research; and data provided by the Swedish authorities.

Deflated by the consumer price index.

Due to the construction of pay agreements, the breakdown is not meaningful in 1987.

In manufacturing.

Derived as hourly wage costs in manufacturing minus real output per manhour.

Table A7.

Sweden: Saving-Investment Balances

(In percent of GDP)

article image
Sources: National Institute of Economic Research; and data provided by the Swedish authorities.
Table A8.

Sweden: General Government Accounts 1/

article image
Sources: National Institute of Economic Research; and Ministry of Finance.

National accounts basis. Excluding nonfinancial public enterprises.

Table A9.

Sweden: The Tax Ratio

(In percent of GDP)

article image
Source: Ministry of Finance, Sweden’s Economy.
Table A10.

Sweden: General Government Revenue and Expenditure in Real Terms 1/

(Annual change in volume; in percent)

article image
Sources: National Institute of Economic Research; Ministry of Finance; and staff calculations.

The private consumption deflator was used to deflate revenues and transfers to households; the GDP deflator was used to deflate other transfers and interest expenditures.

Table A11.

Sweden: Selected Interest Rates

(Period averages, in percent)

article image
Sources: The Riksbank, Quarterly Review; and IMS, International Financial Statistics.

Bonds with five years to maturity.

Three-month Treasury discount notes minus three-month Interbank rate (Germany).

Table A12.

Sweden: Selected Exchange Rates

(Period averages)

article image
Source: IMF, International Financial Statistics.
Table A13.

Sweden: Measures of Liquidity

(Annual percentage change)

article image
Source: The Riksbank.
Table A14.

Sweden: Balance of Payments

(In billions of Swedish kronor)

article image
Sources: The Riksbank; and staff calculations.

Growth in volume of non-oil imports of Swedish partner countries, export weighted; estimated by the IMF Research Department.

Table A15.

Sweden: Competitiveness Indicators

(Annual percentage change)

article image
Sources: IMF, International Financial Statistics; National Institute of Economic Research, The Swedish Economy; and staff estimates.

Total market growth defined as growth in the volume of non-oil imports of Swedish partner countries, export weighted; estimated by the IMF Research Department.

Depreciation is indicated by a decline; IMF Research Department estimates.

Table A16.

Sweden: Direction of Trade

(In percent of total)

article image
Sources: IMF, Direction of Trade Statistics’, and staff calculations.

Figures for Belgium include trade with Luxembourg.