This Selected Issues paper for Ireland highlights that fiscal consolidation resulted in a tremendous reduction in public debt from nearly 100 percent of GDP in 1991 to about 30 percent in 2004. This has reflected a combination of policy decisions and economic circumstances. Excluding 2001, when the economy has been affected by the global economic slowdown, Ireland has in general consistently enjoyed favorable surprises in its public finances. Indeed, during this period, the actual fiscal outturns have exceeded budget forecasts on average by 0.3 percent of GDP a year.


This Selected Issues paper for Ireland highlights that fiscal consolidation resulted in a tremendous reduction in public debt from nearly 100 percent of GDP in 1991 to about 30 percent in 2004. This has reflected a combination of policy decisions and economic circumstances. Excluding 2001, when the economy has been affected by the global economic slowdown, Ireland has in general consistently enjoyed favorable surprises in its public finances. Indeed, during this period, the actual fiscal outturns have exceeded budget forecasts on average by 0.3 percent of GDP a year.

II. The Evolution of Unemployment in Ireland: The Role of Labor Market Policies8

A. Introduction

29. The unemployment rate in Ireland increased dramatically in the early 1980s and remained one of the highest in Europe until the mid-1990s. Since then it has fallen to one of the lowest levels in Europe (Figure 1). The improvement has been broad based. Long-term unemployment, which stood above 10 percent at the end of the 1980s, dropped to 1.5 percent in the 2000s. The rate of unemployment declined for all age groups and education levels. These favorable developments are typically attributed to the upturn of the business cycle, driven by large FDI flows, favorable exchange rate developments, and low interest rates. This paper evaluates whether good labor market policies have also contributed to the sharp fall in unemployment.

Figure 1.
Figure 1.

Ireland: Changes in the Unemployment Rate in Ireland

Citation: IMF Staff Country Reports 2005, 370; 10.5089/9781451818826.002.A002


Long-Term and Short-Term Unemployment Rate

Citation: IMF Staff Country Reports 2005, 370; 10.5089/9781451818826.002.A002

Source: CSO, Household Quarterly Survey and Labor Force Survey. Long-term unemployed are defined as people out of a job for more than a year.

30. The structure of the paper is the following. First, changes in unemployment are decomposed into a cyclical and a structural component. Then follows a discussion of major developments in Irish labor market policies that can induce variations in the structural unemployment rate over time. Finally, the contribution of changes in the tax and benefit system and in labor market institutions to the evolution of structural unemployment is estimated.

B. Cyclical Versus Structural Unemployment

31. Ireland’s economic fortunes changed significantly over the last two decades. The 1980s were characterized by relatively slow growth attributed to negative external shocks and poor macroeconomic policies. In contrast, the second half of the 1990s saw a virtuous cycle of positive shocks, supportive macroeconomic policies, and strong economic growth. That has led some observers to attribute the decline in unemployment mainly to changing cyclical conditions.9

32. In order to isolate the effect of the cycle on unemployment, a time-varying NAIRU has been estimated. NAIRU (“the non-accelerating inflation rate of unemployment”) is the unemployment rate at which there is no tendency for the rate of inflation to change due to wage pressures. The intuition behind the concept is that demand shocks cause cyclical variations in unemployment, but do not have a long-lasting effect on the rate of structural unemployment. Negative demand shocks increase unemployment on impact. That eventually eases wage pressures, bringing about wage moderation compatible with a return of unemployment to its equilibrium level. The estimated gap between the actual unemployment and the NAIRU captures the effect of demand shocks and can be thought of as the cyclical component of unemployment.

33. The estimation follows the methodology of Boone et al. (2002). The model incorporates a two-way causality between unemployment and inflation. The gap between the NAIRU and the actual unemployment rate is assumed to affect the future path of inflation, and policy-induced changes in the inflation rate affect the unemployment gap. Using reasonable parameters for the variance of the unemployment gap relative to the variance of changes in the NAIRU, we find that improving cyclical conditions can explain about 40 percent of the decrease in total unemployment since the second half of the 1980s.10

34. Therefore, at least half of the decline in unemployment can be attributed to changes in the structural unemployment rate. Movements in the structural rate over time are driven by changes in policies and institutions that affect the responsiveness of wages to fluctuations in unemployment, influence the incentives to search for a job, or improve the job-matching process. Empirical and theoretical studies of the driving forces of unemployment have identified the following institutions as important:11

  • The system of wage determination Generally, higher unionization rate is expected to increase the negotiated wage between employers and employees, reducing the equilibrium employment. However, this effect could be offset (or reinforced) by the degree of bargaining coordination. When bargaining is highly coordinated, the aggregate employment implications of wage determination are taken into account, and the wage and employment outcomes would be similar to the social planner’s solution.12

  • The generosity of the unemployment benefit system affects the incentives of unemployed individuals to search for a job. Measures of generosity include the benefit replacement rates and the duration of benefits.

  • Changes in the strictness of enforcement of eligibility for benefits would also affect the equilibrium level of unemployment.

  • The structure of labor taxation and its interaction with the benefit system An increase in labor taxes increases labor costs and/or decreases the net take-home wage. Unless changes in taxes are passed on completely to the workers, equilibrium employment would decline.13 In addition, if benefits are indexed to the gross wage and are not taxed, an increase in taxes would lead to an increase of the effective net replacement rate and a rise of equilibrium unemployment (the reverse would happen if labor taxes are reduced).

  • The introduction (or expansion) of active labor market programs could facilitate job-matching and improve the employability of the long-term unemployed through training and counseling.

C. The Evolution of Labor Market Institutions in Ireland

35. After an unsuccessful experiment with centralized wage bargaining in the 1970s, wage negotiations were effectively decentralized in the early 1980s. In 1987, Ireland introduced an economy-wide system of wage bargaining in an environment of poor economic growth, soaring unemployment, high public debt and a heavy tax burden. Against the background, all social partners agreed that the overriding common objective was to restore sound macroeconomic fundamentals.14 The partners established that this would require improving competitiveness through wage moderation and strengthening the fiscal position to allow future tax reductions. Indeed, a period of relative wage moderation followed, resulting in significant job creation (see Blanchard, 2002). Unions density was very high in the 1970s, but membership declined significantly over the last two decades. Nonetheless, the union’s role in the economy has strengthened after the return of coordinated wage agreements.


Union Density

Citation: IMF Staff Country Reports 2005, 370; 10.5089/9781451818826.002.A002

36. After rising sharply in the 1980s, taxes on labor declined steadily in the 1990s. Both marginal and average tax rates on labor income have been reduced. The structure of taxation also changed. Rising personal exemptions have increased the share of employees that do not have to pay income tax at all, benefiting low-income people. A gradual move towards individualization of income taxation has reduced the tax burden of the second earner in married couples. All these measures have improved the incentives for the economically inactive to join the labor force and for the unemployed to look for a job.


Tax Wedge

Citation: IMF Staff Country Reports 2005, 370; 10.5089/9781451818826.002.A002

37. The generosity of unemployment benefits measured by the gross replacement rate increased in the 1970s and 1980s. At the same time, marginal taxes rates on income were raised and reached very high levels even for low-income earners. The lack of coordination between tax and welfare policies reduced the gap between benefits and the net expected income from work. Based on OECD estimates, the unemployment assistance for married couples without children increased by 56 percent in real terms between 1979 and 1994, although the net income of a worker earning the average manufacturing wage remained constant over the same period. This process has been partially reversed in the 1990s. Although the average gross replacement rate has remained broadly unchanged, the steady reduction of marginal tax rates has resulted in declining net replacement rates, especially for the low-paid. The net replacement rate averaged over three family types earning the average production wage has declined from about 52 percent in 1995 to 46 percent in 2002.

38. In the past, the loss of supplementary benefits—housing subsidies, child benefits, and free medical care—also influenced the incentives to search for employment. The combination of various benefits led to net replacement rates exceeding 100 percent for some low-skilled workers in the 1980s. This problem has been partially addressed through a series of policy measures. A “back-to-work” allowance, introduced in 1993, allowed the long-term unemployed to continue receiving 75 percent of their social welfare payments in the first year of employment, 50 percent in the second, and 25 percent in the third year. Additional measures have been introduced more recently—the rent supplement is no longer withdrawn on taking up employment,15 and child benefits have been increased and uncoupled from unemployment benefits.

39. The enforcement of eligibility for unemployment benefits has been tightened since 1998. Unemployment benefits have effectively unlimited duration in Ireland, so strict enforcement of eligibility is important to keep long-term unemployment in check.16 As part of the National Employment Action Plan, interviews were introduced for those unemployed for six months or more to assess whether they can be matched to existing vacancies or need any specific training. The interviews were progressively phased in for different age groups between 1998 and 2000 and were judged to be very successful in reducing the number of registered unemployed (OECD, 1999).

40. However, in a number of countries that have adopted similar proactive polices to enforcing eligibility for unemployment benefits (Canada, U.S., and New Zealand, for example), an increase in claims for other social benefits has been observed.17 A casual look at changes in the number of social welfare recipients by category suggests that this could be a potential problem for Ireland as well. The decline in unemployment claims between 1998 and 2003 has been more than offset by a rise in claims for other types of social welfare payments. Of course, there may be other reasons for the observed changes, but some analysis of the channels of exit from unemployment may be warranted, and the eligibility for all types of social welfare support should be strictly enforced.

Social Welfare Recipients by Payment Category, 1998–2003

article image
Source: The Irish Labor Market Review, 2004, FAS

41. Spending on active labor market policies in Ireland is relatively high. Active labor market policies include training, public employment services and administration, and subsidized employment (including direct job creation). Evaluations done by the OECD suggest that the different types of active labor market programs vary in their effectiveness. Job-matching services and specific training, for example, are judged to be successful in reducing unemployment; while direct job creation may have limited influence on unemployment in the long run (see Martin, 2000). An increasing share of overall spending in Ireland has gone to public jobs creation. The community employment scheme is the largest program for direct job creation and accounted for about 1.2 percent of employment in 2004 (down from about 3 percent in the mid-1990s). The effectiveness of such spending could be reviewed, especially given the recent tightness of the labor market in Ireland.


Expenditure on Active Labor Market Policies, 2001

Citation: IMF Staff Country Reports 2005, 370; 10.5089/9781451818826.002.A002

Source: OECD, Labor Market Statistics

Active Labor Market Policies Expenditure in Ireland

Citation: IMF Staff Country Reports 2005, 370; 10.5089/9781451818826.002.A002

Source: OECD

42. Not all labor market policies have changed in the direction of greater labor market flexibility. A statutory minimum wage was introduced in 2000 (at about 55 percent of average industrial earnings). However, recent analysis suggests that the effects of the minimum wage of the economy has been marginal since it is binding only for a small share of the work force (Nolan et al., 2003). Progress in product market deregulation has also been uneven.18 In addition, there are two social programs that may serve as a route from unemployment to early retirement and discourage the labor force participation of elderly workers—the pre-retirement allowance and the retirement pension. People over 55 that are unemployed and satisfy means-tested criteria are eligible for a pre-retirement allowance. To be eligible for a retirement pension, a person has to stop working at (or before) age 65, while the regular old-age pension can be received from age 66, independent of employment status.

D. Empirical Evaluation of the Role of Institutions

43. The empirical study of the effect of institutions on structural unemployment is based on a model of the labor market described in Blanchard and Katz (1999).19 In that model, the wage is determined in negotiation between the unions and employers, and firms choose employment at the given wage. Structural unemployment is affected by factors that shift the “wage setting” curve or affect the responsiveness of wages to changes in unemployment—such as changes in labor market polices and institutions. The effect of institutions is typically modeled in a linear form:


where Ut* is the structural rate of unemployment at time t and Xj is a vector of institutional variables. The time varying NAIRU described previously is used as a proxy for the structural rate of unemployment. The independent variables are the gross unemployment benefit replacement rate, the labor tax wedge, union density, an index of bargaining coordination, and a dummy variable for the strictness of benefit eligibility enforcement (takes a value of 1 starting in 1998). For definitions and sources of the data see Annex 1. For ease of interpretation, all variables are in percentage points (except for the index of bargaining coordination and the enforcement dummy). Lack of sufficiently long series prevents the inclusion of active labor market expenditure among the repressors. The equation is estimated by OLS and the results are in the text table.20

44. The results show that changes in labor market policies explain well the changes in the structural rate of unemployment over time. All variables are significant and enter with the expected signs (except for union density). Specifically, ten percentage points reduction in the benefit replacement rate is expected reduce unemployment by about 3.6 percentage points. The elasticity of unemployment with respect to the tax wedge is similar.21 The degree of bargaining coordination has a significant negative effect on unemployment, confirming the general belief that the partnership agreements have contributed to the successful performance of the labor market in the 1990s. The increase in enforcement efforts starting in 1998 seems to have been very effective in reducing unemployment (subject to possible caveats discussed previously).

article image
Note: Asterisks indicate that the estimated coefficient is significantly different from zero at the 1 percent level.

45. Given the uncertainty surrounding estimates of the structural rate of unemployment, a second specification of the model is estimated as a check of the robustness of the results. In that specification, the actual unemployment rate is used as the dependent variable and various demand shocks are added to the set of explanatory variables (see Nickell et al. for a similar approach). The augmented equation is as follows:


where Z is a vector of demand shocks (the real interest rate and labor productivity growth). Changes in labor productivity growth would affect labor demand and respectively unemployment in the short run. On impact, an increase in the real rate of interest raises the required return on investment and reduces labor demand for a given wage level. Over time, high real interest rates raise the cost of capital and reduce capital accumulation. At a fixed ratio of employment to capital, labor demand would fall. Since the Irish and U.K. labor markets are highly integrated, changes in economic conditions in the U.K. can be expected to influence the domestic labor market. Indeed, net migration has been an important channel of adjustment in Ireland. The U.K. unemployment rate is included in the regression to capture the effect of shifts in labor demand in the U.K. on Irish unemployment.

46. The results confirm that demand shocks—interpreted as changes in the interest rate, labor productivity growth, and U.K. unemployment—can explain about a third of the decline in actual unemployment since the second half of the 1980s. Changes in U.K. labor market conditions prove to be a highly significant determinant of the Irish rate of unemployment (although in recent years the link has weakened somewhat as domestic demand conditions have become more important). The coefficients of the labor market institutions remain significant and with the correct sign (with the exception of the coefficient of bargaining coordination which becomes insignificant).

article image
Note: Asterisks indicate that the estimated coefficient is significantly different from zero at the 5 percent (**) or 10 percent (*) level.

E. Concluding Remarks

47. In summary, improved cyclical conditions can explain between a third and a half of the reduction in unemployment in the 1990s, with the rest accounted for by changes in the structural rate of unemployment. The greater flexibility of the labor force has been induced by favorable changes in labor market institutions and policies. Significant reduction in labor taxes, a decline in the effective benefit replacement ratios, stricter enforcement of eligibility for benefits, and the system of coordinated wage negotiations have all contributed to the activation of the unemployed and can explain well the decline in the structural unemployment rate.

48. Going forward, continued flexibility of the labor market would be necessary to sustain rapid growth. The ability to use tax reductions to stimulate labor supply may be nearing its limit. However, further refinements in labor market policies could help facilitate adjustments to economic shocks in the future. The effectiveness of certain active labor market policies, such as the community employment scheme and other job-creation schemes could be reviewed, especially against the background of emerging labor shortages in various areas in the private sector. The rational for the pre-retirement allowance and retirement pension programs may need to be re-examined as these programs discourage labor participation by elderly workers. Limiting the duration of unemployment assistance could improve the incentives of the unemployed to search actively for jobs. Emphasis on enforcement of eligibility for various social support schemes may allow for better targeting of social spending. Finally, the system of social partnership and flexible immigration policies would continue to play an important role in ensuring wage growth consistent with maintaining competitiveness.


  • Arpaia, A. and Carone, G., 2004, “Do Labor Taxes (and Their Composition) Affect Wages in the Short Run and the Long Run?Economic Paper 216, European Commission.

    • Search Google Scholar
    • Export Citation
  • Autor, D. and Duggan, M., 2003, “The Rise in Disability Recipiency and the Decline in Unemployment,” Working Paper, Center for Retirement Research, Boston College.

    • Search Google Scholar
    • Export Citation
  • Blanchard, O. 2002, Comment of Honohan and Walsh, 2002, Brookings Papers on Economic Activity, No.1, pp. 178.

  • Blanchard, O. and Katz, L., 1999, “What We Know and Do Not Know About the Natural Rate of Unemployment,” Journal of Economic Perspectives.

    • Search Google Scholar
    • Export Citation
  • Boone, L., Julliard M., Laxton, D., and N’Diaye, P., 2002, “How Well Do Alternative Time-Varying Parameter Models of the NAIRU Help Policymakers Forecast Unemployment and Inflation in the OECD Countries?” IMF mimeo (available upon request)

    • Search Google Scholar
    • Export Citation
  • Calmfors, 1993, “Centralization of Wage Bargaining and Macroeconomic Performance - a Survey,” OECD Economic Studies, no. 21, 1993.

  • Debrun, X., 2003, “Unemployment and Labor Market Institutions: Why Reforms Pay Off,” World Economic Outlook, IMF.

  • FAS, 2004, “The Irish Labor Market Review,” Dublin.

  • Gruber, J., 2000, “Disability Insurance Benefits and Labor Supply,” Journal of Political Economy, Vol. 108, no. 6, pp. 11621183.

  • Honohan, P. and Walsh, B., 2002, “Catching Up with the Leaders: The Irish Hare,” Brookings Papers on Economic Activity, No.1, pp. 178.

    • Search Google Scholar
    • Export Citation
  • Layard, R., Nickell, S., and Jackman, R., 1991, Unemployment: Macroeconomic Performance and the Labor Market, Oxford University Press.

  • Martin, J., 2000, “What Works Among Active Labor Market Policies? Evidence from OECD Countries’ Experiences,” mimeo OECD.

  • Nickell, N., Nunziata, L., and Ochtel, W., 2002, “Unemployment in the OECD Since the 1960s: What Do We Know?Bank of England.

  • Nickell, N. and Nunziata, L., 2001, “Labor Market Institutions Database,”

  • Nickell, S. and van Ours., 2000, “The Netherlands and the United Kingdom: A European Unemployment Miracle?” Economic Policy, Vol. 15 (April), pp. 13575.

    • Search Google Scholar
    • Export Citation
  • Nolan B., Williams, J., and Blackwell S., 2003, “New Results on the Impact of the Minimum Wage on Irish Firms,” Quarterly Economic Commentary, ESRI, Dublin.

    • Search Google Scholar
    • Export Citation
  • OECD, 1999, Economic Surveys: Ireland. OECD Paris.

  • Walsh, B., 2003, “When Unemployment Disappears: Ireland in the 1990s,” CESifo Working Paper No. 856.

APPENDIX I Data Sources and Definitions

Unemployment Rate is the standardized unemployment rate (OECD Analytical Database).

Benefit Replacement Rate. The data refers to the first year of unemployment benefits as a percentage of average earnings before tax. The rate is averaged over two earnings levels, three family situations and three durations of unemployment. For further details, see the OECD Jobs Study (1994). The OECD publishes this measure once every two years. The intermediate years are interpolated.

Bargaining Coordination. This is an index between 1 and 3, increasing in the degree of coordination in the wage bargaining process on the employers’ as well as on the unions’ side. Constructed by Nickell and Nunziata and available in their labor market institutions database at

Union Density. The ratio of union members to the labor force. The data from 1960 to 1995 is taken from the Nickell and Nunziata database. The data after 1995 is from the OECD and can be found at

Tax Wedge. For 1979–2002 the tax wedge is estimated by the OECD as the sum of income taxes plus employee social security contributions less cash benefits as percentage of labor costs for the average production worker. This data is available once every two years and the remaining years are interpolated. The data for the tax wedge prior to 1979 is from Nickell and Nunziata (2001).

Enforcement is a dummy variable which takes a value of 1 starting in 1998 (when interviews for the unemployed were conducted for the first time).

Real Interest Rate. The long-term rate on government bonds deflated by the GDP deflator (OECD Analytical Database).

Labor Productivity Growth is calculated at the growth of real GDP per hour worked. Source of average hours worked per person: OECD.


Prepared by Dora Iakova, ext. 35365.


The ratio of the variance of the gap to the variance of changes in the NAIRU is set at five in the estimation shown. Varying this parameter within the limits suggested by Boone et al. gives a range of 35 to 50 percent of the decline in unemployment explained by cyclical conditions. Given the uncertainty in choosing the appropriate degree of smoothness of the estimated NAIRU, an alternative estimation of the cyclical part of the decline in unemployment is presented later as a robustness check. I am grateful to Papa N’Diaye for providing the computer code for the NAIRU estimation.


An inverted-U relationship between the degree of bargaining coordination and the level of unemployment is expected to hold. See Calmfors (1993) for an exposition of this theory and relevant empirical evidence.


Arpaia and Carone (2004) show that both perfect and imperfect competition models generally predict a negative impact of labor taxation on employment.


See Chapter IV in IMF Country Report No. 04/349 for a detailed discussion of the role of social partnership agreements.


The continuation of the rent supplement applies only to those taking up part time work or entering approved employment schemes, and to long-term unemployed accepting employment. These rules could create incentives for the short-term unemployed to keep their working hours low once they enter employment.


The unemployed receive benefits for 15 months and then continue to receive unemployment assistance (means-tested) indefinitely.


See Autor and Duggan (2003) for analysis of the U.S., and Gruber (2000) for analysis of Canada.


See Walsh (2003) for a discussion. Highly competitive product markets are expected to increase employment.


The theoretical structure was originally developed by Layard et al. (1991). See Nickell et al. (2002) for a summary of the extensive empirical work based on it.


Unit root tests confirm that most variables are stationary. Two variables - the structural unemployment rate and the benefit replacement rate - fail the test for stationary at the usual significant levels, but since these variables are always between zero and one, they are included in levels.


Since the benefit replacement rate is measured as percent of gross wage, changes in the tax wedge would determine the net replacement rate, which could explain the relatively high elasticity of unemployment with respect to the tax wedge.

Ireland: Selected Issues
Author: International Monetary Fund