Abstract
Denmark’s public expenditure as a share of GDP is the highest in the OECD. The main difference between Denmark and the median OECD country is the larger amount of social protection expenditure. The public health expenditure of Denmark is the second highest in the OECD. Following years of strong public capital accumulation in facilities as well as in training, education, and research, Denmark’s expenditure on public investment is now low. The composition of Denmark’s expenditures is broadly in line with the high expenditure countries.
Composition of Public Expenditure1
This chapter examines public expenditure in Denmark by means of an international comparison on outlays and outcomes. The main conclusion is that while Denmark’s social outcomes are good, there is a case for increasing the efficiency of some public sector services, notably education and health, either by shaving costs, improving outcomes, or some combination of the two.
A. Introduction
1. Denmark’s public expenditure as a share of GDP is the highest in the OECD.2 The bulk of expenditure is concentrated in government consumption, reflecting high public employment (steady at around ⅓ of total employment over the last three decades) and the OECD’s highest public wage bill in GDP terms.
2. The main difference between Denmark and the median OECD country is the larger amount of social protection expenditure owing to Denmark’s generous welfare system, which entails significant spending on disability and family benefits as well as on labor market policies. But while Denmark’s public expenditure is above average, other OECD economies with a similar composition of expenditure still achieve strong social outcomes with lower shares of public expenditure in percent of GDP.
3. Although Denmark’s public finances are strong—with gross debt well below the Stability and Growth Pact’s 60 percent limit—and public expenditure has been instrumental in securing low inequality and a high level of well-being over the past two decades, this exceptional level of spending raises several concerns.
4. First, by requiring a very high level of taxation, it restricts the scope for short-term policy action, such as the fiscal stimulus presented by the new government in Budget 2012. Second, it potentially complicates the sustainability of public finances in the long run, given fiscal pressures from aging, the downward trend in labor factor productivity growth and the expected decline in oil and gas production in the North Sea. Finally, by requiring a heavy tax level, it introduces distortions which harm productivity and constrain long-term growth.
5. The chapter is organized as follows. Sections B and C focus on the composition of Danish public expenditure and compare it with other countries in the OECD. Section D relates spending to outcomes, also providing an international comparison. Section E reviews recent policy changes which have affected public spending. Finally, Section F draws some preliminary policy conclusions.
B. Public Expenditure in Denmark
6. Denmark sits at the top of OECD countries in terms of the amount of public expenditure as a share of GDP.3 In 2011, public expenditure in Denmark represented 52 percent of GDP compared to the OECD average of 43.2 percent. In terms of the size of social protection (measured on a gross basis in percent of 2008 GDP) 4 Denmark is the sixth largest spender in the OECD. Spending on education is 20 percent higher than the average of the organization (in percent of GDP terms), while spending on old age, health, unemployment and disability benefits are all around 10–15 percent higher. Denmark’s spending on unemployment insurance and active labor market policies (ALMPs) more than twice the OECD average in 2008.
7. However, Denmark is far from alone in having total public expenditure at this level. Several countries exceed the OECD average, while the totality of euro area countries exceed the OECD average on public spending. Thus, in addition to comparing the composition of Danish expenditure to the OECD average, which also contains some moderate spenders, it is useful to compare Denmark’s spending pattern with this group of high expenditure countries.5 To facilitate this comparison, data for these countries in the charts contained in this chapter are highlighted in green, and moderate spenders relative to the OECD average are highlighted in blue.
C. Public Expenditures Compared
8. This section looks into the composition of expenditure to investigate whether there are any major differences between Denmark, high expenditure countries, and the OECD overall. The functional composition of expenditure is compared using COFOG functional data, which decomposes government spending into 10 groupings.
9. The composition of Denmark’s expenditures is broadly in line with the high expenditure countries identified above, and with the OECD average in most areas. The exception is social protection, which falls significantly above of the OECD average. Other differences are also apparent, notably in general public services adjusted for interest expenditure, health, and education (all higher), and economic affairs, housing and public order (all lower).
Size and Composition of Public Expenditure
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Sources: OECD and Fund staff calculations.Size and Composition of Public Expenditure
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Sources: OECD and Fund staff calculations.Size and Composition of Public Expenditure
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Sources: OECD and Fund staff calculations.10. Most social protection spending goes toward old-age pensions. However, in Denmark this is marginally lower than in high expenditure countries, and is not strongly out of line with the OECD average. The main difference in social protection spending between Denmark and the OECD average lies actually in spending on unemployment, active labor market programs, family benefit payments and incapacity payments.
11. The high share of spending on old-age pensions is in part explainable in terms of demographics. Comparing old-age pension spending to dependency ratios across the OECD, Denmark sits above the trend line, implying that the pension system is not more generous than in countries which much more rapid aging dynamics (for example, Italy).
12. Denmark’s public health expenditure as a share of GDP is the second highest in the OECD (after Iceland). The majority of health expenditure occurs in the final years of life. Thus, Denmark’s elderly population—above average in numbers in the OECD relative to total population—provides a demographic burden, implying higher-than-average health expenditure. Adjusting expenditure as a share of GDP for the age-spending-profile-weighted-demographic factors, based on estimates by Hagist and Kotlikoff (2005)6 decreases Denmark’s health spending per capita. The total amount of health expenditure also needs to account for private funding which is relatively low but still non-zero in Denmark. On this demographically-adjusted measure of health expenditure, Denmark moves from amongst the highest to being around average as a share of GDP in the OECD (Figure 2).
Public and Private Spending in Key Functional Areas
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Sources: OECD and Fund staff calculations.Public and Private Spending in Key Functional Areas
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Sources: OECD and Fund staff calculations.Public and Private Spending in Key Functional Areas
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Sources: OECD and Fund staff calculations.13. Denmark’s public education spending is amongst the highest in the OECD. This reflects large subsidies to the educational system, which make the system virtually free of charge for its users. And public spending on education is supplemented by private expenditure. However, education expenditure is heavily weighted towards the young, with fewer Danes having higher or further education than peer populations. Since the share of young is relatively low in proportion of Denmark’s population, this helps offset some of the demographic deficit from high pension and health spending. Adjusting education expenditure to account for these factors moves Denmark’s ranking down to around the OECD average as a share of GDP (Figure 2).
14. Finally, following years of strong public capital accumulation in facilities as well as in training, education and research, Denmark’s expenditure on public investment is now low relative to the OECD (Figure 2) and it is expected to remain below 2 percent of GDP until 2020 and beyond in line with Denmark’s 2012 Convergence Program.7
D. Social Outcomes
15. Denmark’s social expenditures are broadly similar to the rest of the OECD once demographic factors are accounted for, but concerns remain over the quality and level of public services.
16. One way to assess this is to compare welfare outcomes to the rest of the OECD, and compare that to the amount of spending to gain a sense of effectiveness. To account for variances in GDP per capita levels across the OECD, the level of spending is stated in US$ per capita on a PPP basis, rather than as a share of GDP as described above. Where relevant—such as in health and education—these spending measures are adjusted for demographic factors and for levels of private expenditure. Points to the north-west in the scatter plot charts (Figure 3) indicate a country with more effective spending. Of course, how effective and efficient public spending is depends on many factors (like population characteristics, past expenditure, income and education levels and immigration rates) which this graphic presentation does not control for. Furthermore, the direction of causality is not always clear—high levels of spending could be a response to poor initial outcomes. Finally, social outcomes of public spending are the result of several if not many years of public spending. Therefore, ideally, they should be evaluated in the context of longer relative spending dynamics, not just of spending differences at one point in time.
Comparison of Social Outcomes to Spending
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Source: OECD and Fund staff calculations.Comparison of Social Outcomes to Spending
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Source: OECD and Fund staff calculations.Comparison of Social Outcomes to Spending
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Source: OECD and Fund staff calculations.17. Denmark’s health outcomes are not very strong. Denmark’s infant mortality rate is close to the moderate spenders than to the higher spenders in the OECD average, while its life expectancy at birth is below the OECD average (Figure 3), having gained only 6.6 years in life expectancy at birth since 1960 relative to 11.2 gained by the average OECD country. Denmark underperforms other high expenditure countries both in terms of infant mortality, and in terms of life expectancy.
18. Importantly, Denmark’s health expenditures appear not to be highly effective, lying away from the efficient frontier of the OECD. Part of this is likely due to Denmark’s non-medical determinants of health, such as high alcohol consumption, a poor diet and past high numbers of smokers among current generations, which are reflected in lifestyle diseases: with growing rates of obesity and diabetes, and high cancer8 and stroke mortality rates.
19. This conclusion is supported by data on Denmark’s health outputs, which are more directly related to spending inputs than the outcomes. Denmark also has an above average doctor-to-population ratio, indicating a highly human-capital-intensive labor input, although the nurse-to-population ratio is below average. Capital-intensive inputs, measured by outputs such as CT and MRI scanners are also relatively high. On the other hand, the number of hospital beds in proportion of the population is relatively low and the beds’ utilization rates (measured by the average length of stay in hospitals for acute care) is the highest in the OECD, indicating—in that case—an efficient use of capital inputs.9 Finally, the quality of care is relatively high, as indicated by very low hospital mortality rate of myocardial infarction (AMI).10
OECD - Income Inequality
(Gini coefficient)
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
OECD - Income Inequality
(Gini coefficient)
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
OECD - Income Inequality
(Gini coefficient)
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
20. Denmark’s education outcomes are slightly better. Denmark’s upper secondary graduation rates are above the OECD average. Yet, these high graduation rates do not always translate into university-level education where Denmark’s enrollment is below the OECD average.
21. Denmark’s PISA score is above the OECD average but not among the highest scores. This is true across the range of subjects covered by the PISA: mathematics, science and reading.
22. These results call into question the effectiveness of education spending in Denmark. Danish annual expenditure per student in US$ terms is well above the OECD average, and it is significantly higher than in countries that achieve much higher results. At the same time, Danish children spend less cumulative time in school (across ages 7–14) than the OECD average. On the other hand, class sizes in Denmark are smaller than the OECD average.
23. A contributing factor in this result is the poor outcomes of immigrants and their children. When looking at the PISA reading scores, most of the discrepancy can be explained by immigrant status, with first-generation immigrant students scoring among the lowest in the OECD, and around 30 percent worse than Danish students with no immigrant background. This is likely due to a number of reasons, such as language skills in Danish, income levels and parent’s education levels: all of which have an impact on education outcomes.
24. Similar patterns can be seen in health indicators (Villadsen et al., 2009). Children of first-generation immigrant women from many developing and emerging market economies have substantially higher fetal and infant mortality compared with the Danish majority population.11 The excess risk is not attributable to socioeconomic conditions. Rather, it is likely due to a number of factors in addition to the provision of health services, including a combination of many other risk factors associated with being part of an immigrant or second generation group, which in turn increases the risk of adverse health outcomes (see Villadsen et al., 2009). These outcomes raise broader economic issues for Denmark, with potentially lower productivity levels of some immigrant and second-generation populations to a large extent determined by education quality and attainment. Thus, this is one area where there is a strong case increasing expenditure.
25. The quality of infrastructure in Denmark is above the OECD average. Denmark scored highly among telecommunication indicators, and on physical infrastructure (Figure 3).
26. Social protection outcomes in Denmark are relatively strong. Denmark has one of the highest levels of equality in the OECD. While the size of redistribution from the tax transfer system (the difference between market and disposable incomes) is around average in Denmark, the highly-equal starting point makes Denmark one of the most equal countries in the OECD, after only Korea, Iceland and Slovenia.
27. Transfer payments typically have a much higher impact on inequality than does the tax system—in the OECD, on average, approximately two-thirds of redistribution occurs through payments. In order to gauge the effectiveness of transfer payments, we compare the amount of redistribution, as measured by the change in the Gini coefficient stemming from those transfers, to the size of the payments, as a share of GDP. The higher the ratio of redistribution to payments, the more higher the efficiency of the transfer payments.
28. The efficiency of Denmark’s transfer payments is above average among those countries for which these data are available. This group can be loosely split between the high tax/transfer European countries, with relatively large universal-style transfer payments, and the (largely Anglophone OECD countries) countries with smaller, more targeted transfer systems. As would be expected with universal systems, the redistribution is less effective, while the targeted (using means testing), and conditional (often to employment) nature of the latter have higher redistributive effectiveness. Denmark’s transfers are typically targeted, but the size of the transfer system is more like the high tax/transfer European countries, with the resulting higher-than-average impact on redistribution from transfers.
E. Recent Policy Changes
29. There have been a number of policy changes directed at containing public spending since 2008. Although these changes will affect spending in the future in a non-negligible way, so far they have neither reduced it nor modified its composition substantially since 2008 and thus they do not compromise the thrust of this analysis. In fact, overall, public expenditure has increased by around 4 percentage points of GDP in Denmark between 2008 and 2011, more than in the rest of the OECD.
30. The three major reforms directed at controlling public spending since 2008 are the introduction of more stringent regulations on public spending, the unemployment insurance reform and the reform of the early retirement system. The reforms are expected to control spending going forward and reduce spending in the longer run, collectively, by around 2–4 percentage points of GDP.
Expenditure rules. These have been modified twice since 2008. Specifically, in 2009, the government introduced sanctions to buttress the enforcement of spending limits for local governments. These have been highly effective, with municipalities actually under-spending relative to the limits both in 2011 and, indicators pointing to under-spending in 2012. Also, expanding upon existing expenditure rules for the sub-national governments, the 2012 Budget Law introduced binding multi-annual expenditure ceilings for each sector of government to be in effect from 2014 onwards. The ceilings set nominal upper limits on actual spending and include a broad range of public expenditures. However, they exclude unemployment-related—and strongly cyclically-sensitive—costs such as unemployment benefits and cash benefits, so that the automatic stabilizers can work in line with cyclical fluctuations. The expenditure ceilings that will be in place from 2014 onward will be underpinned by measures to improve finance management, including through a better coordination of responsibilities between levels of government, as well as economic sanctions, cf. above.
Unemployment Insurance Reform. This reform, implemented in 2010, halves the length of time that workers can claim the benefit from four years to two and doubling the length of time they have to pay in to the system before being eligible from six months to one year. The reform is expected to generate a saving of 1½bn DKK in 2013, and permanent annual savings of around 4½bn DKK starting from 2014. Under the current government, the length of unemployment benefits was temporarily extended for those with expiration in mid 2012 so that the first “accelerated” expirations would take place only in January 2013.12 Relatively to the original proposal, this extension has worsened public spending by around ½bn DKK when taken cumulatively over 2012–2013.
Retirement Reform. This reform is based in the need to ensure sustainability of the pension system in the face of the decreasing size of the Danish workforce and demographic projections. The reform is based upon three key measures: (i) adding a further five years to the gradual rise in pension and early retirement age set out in the 2006 Welfare Agreement; (ii) maintaining the current indexation of pension age and early retirement; (iii) gradually reducing the available period of early retirement from five to three years, by progressively increasing the early retirement age from 62 to 64 years between 2018 and 2023.13 In the near-term the pension reform increased public spending by around 1 percent of GDP, as it resulted in a one-off public expenditure in 2012. In the longer term the reform is set to stimulate GDP growth through an increase in the employment with an effect estimated at 18bn DKK from 2020 onwards. In addition to the proposed adjustments, a senior disability retirement payment has been introduced, which is a variant of the existing disability retirement payment. The senior disability pension is aimed at employees over the age of 62 who cannot continue to work because their job is physically demanding.
Effectiveness of Transfer System
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Effectiveness of Transfer System
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
Effectiveness of Transfer System
Citation: IMF Staff Country Reports 2013, 023; 10.5089/9781475527322.002.A001
F. Concluding Remarks
31. While Denmark’s social outcomes are of a relatively high quality, there is an argument for better targeted health and education expenditure considering that there are a number of more moderate expenditure countries which succeed in delivering high quality social outcomes as well as similar expenditure countries which achieve higher targets with similar amounts of civilian spending. The generousness of Denmark’s safety nets, especially disability benefits aimed at raising participation of certain categories of disabled citizens, could also be revisited against the backdrop of social outcomes. Failure to tackle these problems would lead fiscal sustainability problems.
32. Policy actions taken are moving in the direction of containing public spending, with the introduction in 2012 of mandatory spending ceilings for all levels of government, the reform of unemployment insurance and of early retirement, plus steps to improve the consistency of responsibilities between central and sub-national governments, initiatives to raise labor supply via reductions in the income tax wedge, and improvements in the evaluation framework for both students and staff. However, with its elderly population adding a sizeable and increasing demographic cost, containing public spending while ensuring that it reaches the efficiency frontier is complicated.
33. Exploring additional, specific ways to contain spending while raising outcomes is beyond the scope of this chapter. However, one building block for such exploration could be a comprehensive spending review to identify areas of where public resources could be used more effectively and efficiently. So future policy actions could focus on a review of public spending. Likewise, raising labor supply, especially of older workers and increasing productivity levels of the employed could reduce spending considerably without affecting its efficiency. To increase the effectiveness of spending, Denmark should also look to lessons from those moderate expenditure countries, particularly in the design of policies to increase the degree of targeting and conditionality to provide more bang for the social transfer krone especially on labor market-related benefits, including disability.
Prepared by Nicoletta Batini (nbatini@imf.org) with research assistance provided by Tom Dowling.
Based on the OECD’s most recent Government at a Glance (2011) that, in turn, uses latest available data which are largely 2009 data.
Care must be taken in comparing public expenditure across countries. In Denmark income transfers are usually taxable, while income transfers in some OECD countries are not taxable. This difference does not affect the conclusions of this paper.
2008 is chosen for comparing spending by function of government in Denmark to other countries because this allows a comparison of expenditure before the impact of the global financial crisis. It also marks the date of the most recent OECD’s Classification of Functions of Government (COFOG) data used here.
We define “high expenditure” countries as countries with public spending in percent of GDP above the OECD32 average (including, notably, countries like France, Germany, Iceland, Italy and Sweden). By contrast we define “moderate expenditure” countries with public spending-to-GDP below the OECD32 average (including, notably, Australia, Canada, New Zealand, and the United States).
Laurence Kotlikoff and Christian Hagist, “Who’s Going Broke?” National Bureau of Economic Research, Working Paper No. 11833, December 2005, p. 29.
Note that it is difficult to compare public investments across countries due to differences in how different tasks are organized in the public or private sector. In Denmark, quite a few investments are carried out by publicly-held companies and businesses that are measured as private investments in the National Accounts and are thus not included in public investments (e.g. tunnels and bridge-links, water and sewage and many others), so the figure provided above likely underestimates the true level of public investment in relative terms.
Denmark’s breast and prostate cancer mortality rates, for example, are the highest and second highest in the OECD, respectively.
The average length of stay in hospitals for acute care is often considered a measure of efficiency. All other things being equal, a shorter stay will reduce the cost per discharge and shift care from inpatient to less expensive post-acute settings. However, shorter stays tend to be more service-intensive and costly per day. Too short a stay could even signal a potential for more adverse health outcomes.
AMI case-fatality rate is a good measure of acute care quality because it reflects the processes of care for AMI, such as effective medical interventions including thrombolysis, early treatment with aspirin and beta-blockers, and coordinated and timely transport of patients. AMI case fatality rates have been used for hospital benchmarking in several countries including Canada, Denmark, the United Kingdom and the United States.
Note that the term “ethnicity” and its use in health research is a debated and controversial topic. We refer the reader to other works for a discussion of these issues, for example, Bhopal, R. (2004). “Glossary of terms relating to ethnicity and race: for reflection and debate”, in Journal of Epidemiology and Community Health 2004, Vol. 58:441–5.
Under the budget deal reached in November 2012, for people whose unemployment insurance benefits run out in the first half of 2013, there will be a temporary 6 months option to continue with cash benefits provided one accepts a training course or a period of job training. These benefits will be significantly lower than the standard UI benefits, but in contrary to those, they will not depend on spouse income or on the level of the recipient’s wealth. The additional fiscal cost of this new provision is estimated to be small.
Taken together, these imply that by 2023 an employee will be able to take early retirement between 64 and 67, at which point they join the ordinary pension scheme.
