Denmark
Selected Issues

This Selected Issues paper for Denmark shows that a demographic shift will have several impacts on the economy. The decline in the relative size of the labor force will result in relatively fewer goods produced. This effect will be particularly strong, because a large portion of the current baby-boomers are in their peak earning years, and their retirement will have a dramatic effect on productivity and on overall production. Demark has imposed a restrictive policy requiring that immigrants prove they have a job that meets wage and working condition standards before getting a work permit.

Abstract

This Selected Issues paper for Denmark shows that a demographic shift will have several impacts on the economy. The decline in the relative size of the labor force will result in relatively fewer goods produced. This effect will be particularly strong, because a large portion of the current baby-boomers are in their peak earning years, and their retirement will have a dramatic effect on productivity and on overall production. Demark has imposed a restrictive policy requiring that immigrants prove they have a job that meets wage and working condition standards before getting a work permit.

I. Demographic Changes and Fiscal Sustainability in Denmark1

A. Overview and Conclusions

1. Like most developed countries, Denmark is facing significant demographic changes. Over the next few decades, the over-65 population will jump from 16 percent to 25 percent of the total population. At the same time, the working-age population will shrink from 65 percent to 57 percent. As a consequence, the ratio of working-age to elderly individuals will plummet from 4:1 to just over 2:1.

2. This demographic shift will have several impacts on the economy. First, all else equal, the decline in the relative size of the labor force will result in relatively fewer goods produced. This effect will be particularly strong, because a large portion of the current baby– boomers are in their peak earning years, and their retirement will have a dramatic effect on productivity and on overall production. Second, the increase in the elderly population could have a substantial impact on the demand for goods and services, especially for health services and old–age care. Finally, and most importantly for this paper, the demographic shift is likely to have an important impact on public finances. Much of old–age spending in Denmark is facilitated by the state—either through the provision of public goods (medical services and old–age care) or through transfer payments (primarily pensions). As the population ages and the demand for public spending increases, the government’s intertemporal budget constraint will come under pressure, and taxes may need to rise or per capita spending may have to fall to maintain fiscal sustainability.

3. Moreover, Demark may face additional demographic challenges from immigration. The accession treaties that ushered the new member states into the European Union call for unrestricted labor mobility across all countries by end–2010. The treaties allowed member countries to temporarily place restrictions on inward migration from new member states. Demark has imposed a relatively restrictive policy—requiring that immigrants prove they have a job that meets wage and working condition standards before getting a work permit—which will have to be relaxed. At the same time, Denmark has been quite welcoming to refugees and asylum seekers.

4. Although the economic challenges of ageing have been studied extensively, issues associated with fiscal pressures and sustainability have received relatively little attention. This paper is an attempt to bridge this gap, using a calibrated model to look at the case of Denmark. The underlying model is a Ramsey growth model, where rational, forward– looking individuals produce goods and divide their income between consumption and saving, which adds to the future capital stock. Aggregate consumption is distributed among individuals according to their age–related needs, using private and public transfer schemes. In this model, the public transfer schemes are financed with non–distorting taxes on income, so there are no indirect effects on economic growth.

5. The main conclusions of the paper are as follows:

  • Demographics will have a significant adverse effect on the labor market, on real GDP, and on public finances. In the baseline scenario – which is calibrated on policies in place at the end of 2005 – he fiscal sustainability gap is estimated at about 4 percent of GDP. This is somewhat greater than estimates made by the authorities and by the European Commission (2005). The main differences appear to be related to more optimistic assumptions about health care and old–age care expenditures.

  • The 2006 Welfare Agreement and the 2007 Jobs Plan go a long way towards putting fiscal policy on more solid ground. Together, these policy initiatives improve the fiscal balance by about 4 percentage points by 2005. However, since the Agreement is not fully phased in until 2027, many pension recipients continue to be “grandfathered in”, and a sustainability gap of about 2 percentage points remains.

  • The estimated sustainability gap is larger than estimates by the Ministry of Finance, owing mostly to the Ministry’s assumption of “healthy ageing,” which proposes that a 70-year-old in 2050, for example, will require less health care and old-age care than an 70-year-old today.

  • Recent policy changes have widened the gap a bit, but some recent proposals by the Labor Market Commission would likely close the gap. However, these proposals have not been received warmly by some of Denmark’s social partners.

  • Finally, it is important to recognize that this study is based on many assumptions, and the main conclusions are quite sensitive to some of them.

6. The paper is organized as follows. Section II discusses the demographic changes underway in Denmark and provides a preview of how they will directly affect the supply of labor and the demand for public sector benefits. Section III examines in more detail how the demographic shifts will affect the labor force, real GDP, and public finances. Section IV examines the effects of recent policy changes – including the Welfare Agreement and Jobs Plan, and recent tax and expenditure changes – and discusses the possible effects of the recent proposals made by the Labor Market Commission.

B. Demographic Changes

7. This section reviews recent and projected demographic changes. The focus is on ageing and immigration. An ageing population is a feature of many developed countries, largely driven by decreases in fertility rates and declines in old–age mortality rates. Denmark’s ageing “problem” is not as severe as in some other countries, where fertility rates have dropped further. Indeed, many countries—most notably, Spain, Italy, Greece, Japan, and Germany—are ageing much faster and will end up with significantly larger elderly population shares—see IMF (2001).

8. Immigration interacts with population ageing in several ways. Immigrants, on average, tend to be younger than the average native population. This can be an advantage, if immigrants move quickly into productive employment, providing a boost to the labor force and to a net positive contribution to public finances. On the other hand, immigration can be a burden, if immigrants end up in low-skill, low-wage jobs or are unable or unwilling to participate in the domestic labor market.

Population Ageing

9. Denmark’s population is ageing. As shown in Figure 1, the ratio of working–age to elderly individuals is expected to fall from 4:1 today to just over 2:1 in 2050. The decline reflects a fall in the working–age population (WAP) – currently defined as individuals between 16 to 64 years of age – and a sharp rise in the elderly population – 65 years of age and above.

Figure 1.
Figure 1.

Denmark: Ratio of Working–Age to Elderly Population, 2008–2050

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Statistics Denmark

10. Figure 2 provides a snapshot of this trend. It shows the age distribution for three selected years—2008, 2015, and 2050. The distribution in 2008 (the solid line) shows that the 2008 distribution has several modes; the largest mode corresponds to cohorts in the early 40s, while somewhat smaller clusters of individuals are in their early 60s and in their pre-teens. It is also interesting to note that the distribution drops off sharply for elderly individuals.

Figure 2.
Figure 2.

Denmark: Population Distribution, 2008, 2025, and 2050

(Thousands)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Statistics Denmark

11. As the distribution shifts to the right, from 2008 to 2025, the largest group of cohorts are pushed into their 60s, while the younger group moves into employment, and the older group moves into retirement. The number of individuals in the working-age population (16-64) declines over this period, while the number of 65+ year-olds increases markedly. The share of the elderly in the overall population becomes even larger by 2050, when the largest group of cohorts has moved into old-age.

The Immigrant Population

12. The population is projected to increase by about 360 thousand (a 6½ percent change) from 2008 to 2050, with almost all of the increase expected to come from new immigrants or from the descendants of immigrants. Figure 3 shows the actual and projected population over the 1981–to–2050 period, decomposed into 5 ancestral categories: (i) individuals with Danish origins (born in Denmark from Danish parents), (ii) immigrants from developed countries (DC), (iii) immigrants from less developed counties (LDC), (iv) descendants of immigrants from developed countries, and (v) descendants from immigrants from less developed countries. As the figure illustrates, the population with Danish origins is expected to be relatively stagnant over the next several decades, while the number of immigrants is projected to increase by more than 50 percent, and descendants are projected to more than double (albeit from a much smaller level).

Figure 3.
Figure 3.

Denmark: Composition of Population, by Ancestry, 1981–2050

(Millions)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: DREAM Model

13. Whether these trends in the immigrant population will mitigate or exacerbate the challenges of ageing depends on how the immigrant population differs from the native population. The remainder of this section compares these population segments on the basis of age, educational attainment, employment activity, and net contributions to public finances.2 The comparisons can be summarized as follows:

  • The share of non–natives in the working age population has increased steadily in recent years. As shown in Table 1, the share of immigrants in the WAP has increased almost 4 percentage points from 1991 to 2006. This increase is concentrated among the relatively young—especially among those under 45 years old. In addition, the table indicates that new immigrants have been relatively unskilled, with immigrants pushing up the share of total individuals without a high school education (labeled “Basic”).

Table 1.

Denmark: Change in Importance of Immigrant Population, by Age and Educational Attainment

(Change from 1991 to 2006, percentage points)

article image
Basic=8th grade education or less; GUS=general upper secondary; VUS=vocational upper secondary; VS=vocational school; SHE, MHE, and LHE=short-, medium-, and long-cycle higher education, respectively.Source: Statistics Denmark
  • Non-natives are expected to become a larger share of the working-age population. As seen in Figure 4, from 2008 to 2050, the number of immigrants and descendants in the working age population will increase by more than 185,000 individuals, while the native population falls by more than 400,000. As a result, the share of non-natives in the WAP is projected to increase from about 10 percent in 2008 to about 16 percent in 2050.

  • However, non-natives will also make up a larger share of the elderly population. As illustrated in Figure 5, the elderly population is expected to rise sharply for all ancestral categories, but elderly immigrants from less developed countries are expected to increase more rapidly. As a result, the share of non-natives in the elderly population is seen to increase from 4 percent in 2008 to about 10 percent in 2050.

Figure 4.
Figure 4.

Denmark: Composition of Working-Age Population, by Ancestry, 2008-2050

(Millions)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Statistics Denmark
Figure 5.
Figure 5.

Denmark: Composition of Elderly Population, by Ancestry, 2008-2050

(Millions)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Statistics Denmark
  • Economic activity varies considerably across ancestral categories.3 As illustrated in Figure 6, labor force participation rates (LFPRs) are quite high for native Danes (the solid line), particularly because labor force participation among women is about the same as for men. LFPRs for immigrants from less developed countries are uniformly below those of other categories, reflecting differences in characteristics such as job skills, language, or culture. LFPRs for immigrants from developed countries are lower for younger individuals, but are nearly the same as for native Danes for individuals over 40 years old. Descendants from both developed and less developed countries also have nearly the same participation rates as native Danes for all age groups, suggesting that assimilation in the labor market occurs quickly in Denmark.

Figure 6.
Figure 6.

Denmark: Labor Force Participation Rates, by Age and Ancestry, 2006

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Statistics Denmark
  • Unemployment rates also show a great deal of variation across ancestral categories—Figure 7. The patterns are similar to those for labor force participation rates. Unemployment rates are relatively low for native Danes, and quite high for immigrants from less developed countries. Descendants from lesser developed countries have relatively high unemployment rates when young, but are nearly indistinguishable from other categories for those older than 40 years old. Finally, the unemployment rates for other groups are nearly the same as for native Danes.

Figure 7.
Figure 7.

Denmark: Unemployment Rates, by Age and Ancestry, 2006

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Statistics Denmark
  • Finally, ancestral categories place different demands on the public sector. Figure 8 and 9 show the distribution of public consumption and public transfers, respectively, for men in 3 different ancestral categories.4 For public consumption, immigrants under the age of 30 have a slight tendency to consume smaller amounts of public benefits, which Schou (2005) attributes to lower consumption of education. Differences among the groups are neglible for older individuals. It is also quite striking that public consumption rises drastically after age 70, reflecting the need for medical care and long–term care. Immigrants from less developed countries receive considerably more government transfers, likely reflecting lower participation in the labor market, as discussed earlier. As pointed out by Schou, however, this tendency fades in old age, because fewer immigrants tend to receive public pensions.

Figure 8.
Figure 8.

Denmark: Public Consumption for Men, by Age, 2005

(Thousands of DKR)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: DREAM Model
Figure 9.
Figure 9.

Denmark: Public Transfers for Men, by Age, 2005

(Thousands of DKR)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: DREAM Model

C. When Is Fiscal Policy Sustainable?

14. There are several ways to evaluate whether current fiscal policies are sustainable. All criteria start with an equation describing public debt evolution:

bt+1=(rtγt)btpst

where bt+1 is the debt-to-GDP ratio, rt is the real interest rate, γt is the growth rate of real GDP, and pst is the primary surplus. Repeated substitution of the equation produces the debt ratio for time t+s:

bt+sRt+s=btΣk=0s1pst+kRt+k

where

Rt+s=Πi=0s1(1+it+i+γt+i)

This relationship simply states that the discounted value of future debt—the left-hand side of the equation—is equal to current debt minus the sum of discounted future primary surpluses—the right-hand side of the equation.

15. As discussed in Blanchard et al. (1990), there are many debt paths that satisfy the above intertemporal budget constraint (IBC) over s periods. The most common definition of debt sustainability – and the one used in this paper and by the Danish authorities – is a scenario where the debt target at time t+s is the current debt ratio (bt). However, the IBC also allows for the debt ratio to rise to a higher but stable level (a somewhat less restrictive definition) or to converge to zero at time t+s (a more restrictive definition).5 The definition can be made operational by setting a debt target for time t+s (b¯t+s = bt) and then calculating the change in the primary surplus (äps) in each time period between t and t+s that satisfies the sustainability definition. In other words, äps must satisfy the following equation:

b¯t+sRt+s=btRt+s=btΣk=0s1pst+k+ΔpsRt+k

D. A Baseline Forecast

16. This section of the paper develops a baseline scenario for the macro economy and the fiscal sector in Denmark from 2008 to 2050. The baseline scenario is based on government policies that were in effect at the end of 2005. This counterfactual scenario allows for an analysis of the effects on fiscal sustainability of both past and prospective policies: (a) the 2006 Welfare Agreement and the 2007 Jobs Plan, (b) recent changes in tax and expenditure policies, and (c) recent proposals by the Labor Market Commission.

17. The model’s production function has three inputs – technology, labor, and capital. Technology is assumed to evolve exogenously. The capital stock is endogenous, and adjusts optimally to demographic changes. The labor force is also exogenous, based on Danish official forecasts. A key feature of the model in this paper, however, the labor force in this model is disaggregated by age, educational attainment, and ancestry. This allows for a rich analysis of the effects of demographic changes on the labor force and, therefore, on real GDP.

The Effective Labor Force

18. Goods are produced at time t using capital, labor, and labor–augmenting technology as follows:

Yt =Ktα (EtLt)1α

where Yt is real output, Kt is the real capital stock, Et is the average level of labor efficiency, Lt is demographically-weighted level of employment, and α represents capital’s share of output.

19. The demographically–weighted labor force is constructed as follows:6

Lt=Σi=110Σj=17Σk=15ωijkNijkt

where Nijkt is the population at time t, disaggregated by:

  • age(i) = 16-19, 20-24, 25-29, …. 60-64;

  • education (j) = basic, general upper secondary, vocational upper secondary, vocational, short-cycle higher, medium-cycle higher, and long-cycle higher; and

  • ancestry (k) = Danish origin, immigrant from a developed country, immigrant from a less developed country, descendant from a developed country, and descendant from a less developed country;

and where population segments are weighted (ωijk) to reflect differences (across age, education, and ancestry) in labor force participation, unemployment rate, and productivity. Specifically, the demographic weights are calculated as:

ωijk=LFPRik*(1URik)*Wij

where LFPRik is the labor force participation rate (across age and ancestry groups), URik the unemployment rate (across age and ancestry groups), and Wij is the average hourly wage (across age and education groups).

20. Note that three important assumptions are made based on available data:

  • The average hourly wage rate is a good proxy for relative productivity;

  • Individuals across immigration categories (k) earn the same wage, after adjusting for age (i) and education (j);

  • Individuals across educational categories (j) have the same labor force participation and unemployment rates, after adjusting for age (i) and ancestry (k).

In addition, it is assumed that average labor efficiency (Et) grows 2.2 percent per year, which Et is consistent with the average growth rate during the 1 980s and 1 990s but somewhat faster than the average growth rate in recent years.7

21. Figure 10 shows three measures of the growth rate of the projected Danish labor force through 2050. The solid line shows the growth rate of the working-age population.

Figure 10.
Figure 10.

Denmark: Various Measures of the Labor Force, 2008–2050

(Percent change)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: IMF staff calculations.

The short-dashed line shows the growth rate of employment—the working-age population adjusted for differences in labor force participation and unemployment rates across the demographic categories. Finally, the long–dashed line plots the growth rate of employment adjusted for demographic differences in productivity, as proxied by earnings. The figure indicates that both types of adjustment are important in Denmark: Through 2030, the increased importance of immigrants—who tend to have lower rates of economic activity and lower earnings – and the large shift into retirement of middle–aged people—which tend to have relatively high earning power—will significantly and adversely affect the effective labor force. After 2030, when immigration ratios and relative productivity stabilizes, the demographically–adjusted employment growth rate rises rapidly and then levels off.

The Capital Stock and Real GDP

22. The final step in calculating real GDP is to forecast the real capital stock. The growth model predicts that the capital stock grows in line with the effective labor force. More specifically, the equilibrium capital-to-effective-labor ratio is:

KEtLt=[KY]1/(1α)=[(1τk)αr+δ]1/(1α)

where τK is the average tax rate on capital income, r is the real interest rate, and δ is the depreciation rate of capital.

23. This equilibrium relationship suggests two approaches for pinning down the optimal capital–to–effective–labor ratio:

  • The capital-to-output ratio has fluctuated between 2.8 and 2.9 over the last 25 years, averaging 2.83. Using the formula above, this implies that the optimal capital-to-effective-labor ratio should be about 6.6.

  • The average effective tax rate on capital is about 30 percent.8 Using national accounts data, á and ä are estimated to be about 0.45 and 0.045, respectively. The real return on capital is assumed to be about 6 percent. Using the formula above, the optimal capital-to-effective-labor ratio should be about 7.4.

  • This first calibration is likely more accurate, since more parameters are directly observable. In the baseline scenario, I will assume that the long-run ratio is 6.6 for Denmark. As it turns out, the qualitative nature of the conclusions are not affected if a higher ratio is used.

24. An important implication of this analysis is that the growth rate of the capital stock should fall somewhat over the next several years, since the effective labor force is expected to also slow over that time period. This paper assumes that the capital stock adjusts very slowly to these changes, consistent with adjustment rates presented in Barro and Sala-i-Martin (2004).

25. Using the estimates for the effective labor force in the previous section and the assumptions for the capital stock above, it is straightforward to compute estimates of GDP using the production function discussed above. Figure 11 shows the outcome of this exercise. Over the next few years, GDP (the bold line) is projected to grow at a 1¼ percent rate, much slower than in recent years, owing to a slowdown in the effective labor force (the dashed line). However, after 2030, the effects of the ageing population start to wane and the immigrant population stabilizes. As a consequence, real GDP growth rebounds toward the end of the forecast horizon.

Figure 11.
Figure 11.

Denmark: Real GDP and Effective Labor Force, 2008–2050

(Percent change)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: IMF staff

26. This baseline forecast for real GDP is subject to a number of uncertainties. First, labor efficiency has been growing very slowing in recent years. Most analysts seem to view this as a largely cyclical phenomenon, and this paper assumes that efficiency growth will rebound after the current downturn is over. If efficiency growth continues to languish, however, then the estimates shown in Figure 11 will be too high. Finally, there is a great deal of uncertainty regarding the projected inflow of immigrants after Denmark lifts restrictions on immigration from new member states, in line with EU requirements.

A Baseline Forecast of the Fiscal Accounts

27. Finally, this section makes some rough calculations of future public sector revenues and expenditures. The focus is on calculating the direct effects of demographic changes on public expenditures. If public expenditures (as a percent of GDP) are expected to rise over time in an unsustainable way, then either average tax rates will also have to increase or benefit levels will need to be lowered, in order to satisfy the government’s intertemporal budget constraint (IBC). This will lead to indirect effects. If tax rates on capital are increased and capital is relatively mobile, then capital accumulation is likely to slow sharply. Tax bases would also slow, requiring even higher taxes to satisfy the IBC. Alternatively, if tax rates on labor are increased and labor is relatively immobile, then the effects on output could be small or even negligible.9

28. Nominal revenues are assumed to be constant over time, as a percent of nominal GDP. This is the same assumption made by the Danish authorities, and it is consistent with the thrust of the government’s aim to freeze taxes, on average, at least through 2015.

29. Nominal expenditures for spending category s for year t (Xst) can be written as:

Xst=Pst*ρst*Σi=110Σk=15ΨsikNikt(t2008)

where Pit denotes the nominal price index for spending category s at time t relative to some base year; ρst is the real cost index of providing public services in category s for time t relative to the base year; ψst represents per capita real expenditures in category s for persons in demographic categories i and k (age and ancestry) for a given base year; and is the Nikt number of persons in demographic categories i and k at time t.

30. This equation suggests a simple, straightforward way of projecting fiscal expenditures, using available data:

  • The first step is to use data on per capita expenditures (ψsik) for spending category s and for various age (i) and ancestral (j) categories for 2005 – illustrated in Figures 7 and 8 –and multiply by the number of persons (Nikt) in each of the demographic categories over the forecast horizon.

  • Summing up over all demographic categories and multiplying by the real costs of these expenditures in a given year (ρst) yields an estimate of total real expenditures for spending category s in year t.

  • Finally, nominal expenditures (Xst) are calculated by multiplying real expenditures by the appropriate price index (Pst).

It is also assumed that the real cost of all spending categories (ρst) rises with the level of overall labor productivity (real wages) in the total economy and that Pst grows over time at the rate of overall inflation. The specific assumptions are:

  • Inflation is expected to be 3.1 percent in 2009, 2.3 percent in 2010, 2.1 percent in 2010 and 2011, and 2 percent thereafter; and

  • Labor productivity growth is assumed to grow at a constant rate of 1.8 percent over the forecast horizon.

Nominal GDP is simply real GDP times the nominal price index. Note that assumptions for labor productivity and overall inflation are the same as above.

31. Table 2 presents the outcome of this simple projection exercise. Overall, total government expenditures increase by about 15 percentage points of GDP during 2018–2050, much more than the official estimates provided to EC committee on ageing (about 3 ½ percentage points of GDP). The largest contributors to this increase are health care (4.9 percentage points) and pension expenditures (about 4 percentage points). Increases in “other social spending” and “other transfers” account for the rest of the increase. The baseline scenario also indicates a 3.9 percentage point gap in fiscal sustainability. That is, there would need to be an increase in the fiscal balance of 3.9 percent of GDP in each period from 2008 to 2050 to keep the debt to GDP ratio at the 2008 level.

Table 2.

Denmark: Baseline Projections of the Structural Balance, 2008–2050

(Percent of GDP)

article image
Source: IMF calculations.

32. A number of underlying assumptions need to be studied more carefully. First, health care costs could be over- or under-estimated. On one hand, many countries, including Denmark, are experiencing sharp increases in relative prices for health care. If demand pressures continue to put upward pressure on the costs of health care, the overall costs of ageing would be higher than shown in Table 2. On the other hand, these calculations could be biased upwards, since I assume that the health of a 70-year-old in 2040—and her associated health care costs—will be the same as a 70-year old in 2005. Education costs could be understated, if more young people seek education to be employable in a high-tech, competitive global economy. Also, if workers will be spending more years in the workplace, they might need more education later in life, in response to sectoral economic shifts.

33. Finally, as mentioned earlier, these calculations do not account for the effects that higher taxes might have on the labor-leisure choice or the savings-consumption choice. Indeed, higher government expenditures due to ageing are likely to lead to higher taxes, lower employment rates, slower capital accumulation, and slower economic growth. In this case depending on tax policy choices, ageing costs as a percent of GDP could be even higher than the calculations in Table 2 suggest.

E. Effects of Recent Policy Changes and Proposals

34. This section of the paper examine the effects on fiscal sustainability of two recent changes in fiscal policy – the 2006 Welfare Agreement and some recent changes in tax policy – and a recent proposal by the Labor Market Commission to bring forward the increases in age limits that are in the Welfare Agreement.

The Effects of the Welfare Agreement

35. The most important features of the Agreement are an initial increase in retirement ages, and, thereafter, an indexation of retirement ages to longevity. The current retirement age for voluntary early retirement pensions (VERP) and old-age pensions (OAP) is 60 and 65, respectively. Given current life expectancies, a new pensioner in 2008 can expect to spend between 17 to 22 years in retirement, significantly longer than previous cohorts. Moreover, life expectancies are projected to increase further—about ½ of a year every 10 years, putting substantial pressures on pension expenditures and effectively reducing the relative size of the working-age population. The Welfare Agreement raises the age limit for VERPs from 60 to 62 years starting in 2019, and the age limit is then indexed to the life expectancy of 60 year-olds thereafter, according to a specified formula. The age limit for OAPs will increase from 65 to 67 starting in 2024, and, thereafter, will be the VERP age limit plus 5 years.

36. These changes have a dramatic effect on the old-age dependency ratio (Figure 12), both increasing the number of individuals in the work force and by reducing the number of pensioners. In terms of the economic model, there are a number of important effects. First, the labor force in each period is augmented by the number of individuals that would have been in retirement without the Welfare Agreement requirements. Although this has the effect of increasing various tax bases, I assume that these increases have no distributional effects, and revenues as a percent of GDP are unchanged. Second, the reduction in the number of pensioners reduces transfer incomes. As a consequence, income transfers as a percent of GDP fall because the level of transfers will decrease (the numerator) and the level of GDP will increase (the denominator).

Figure 12.
Figure 12.

Denmark: The Effects of the Welfare Agreement on the Old-Age Dependency Ratio

(Mandatory pensioners as a percent of working-age population)

Citation: IMF Staff Country Reports 2008, 380; 10.5089/9781451811247.002.A001

Source: Ministry of Finance and IMF staff

37. As shown in Figures 6 and 7, labor force participation rates and employment rates drop sharply as individuals get close to retirement age. The 2007 Jobs Plan strengthens work incentives for individuals as they approach retirement, offering tax incentives to stay in full time work and some retirement benefits for those who shift from full- to part-time work. The authorities (2007 Convergence Program) estimate that the combined effects of the Welfare Agreement and the Jobs Plan will not only add more workers to the labor force, but their employment rate and average hours worked will increase substantially.

38. In order to simulate the effects of these two policy changes, I make the following adjustments to the baseline scenario:

  • The labor force will increase significantly following the phasing in of the Welfare Agreement in 2019, as the age limits on retirement are gradually increased. All affected individuals are assumed to be pushed into the labor market as a result of the policy change. In reality, there will likely be some slippages, as some people elect to be out of the labor force or move into sickness or disability schemes.

  • In addition, I also assume that the employment prospects of older workers will be improved by the Jobs Plan. In particular, labor force participation rates and employment rates for workers 60 to 70 years old are assumed to rise gradually, starting in 2009 and eventually yielding job performance on par with 55 to 60 year olds in 2008.

One important difference between these assumptions and those made by the Ministry of Finance is that they not include any effects from “healthy ageing”. In contrast, the Ministry of Finance (and some other studies) assume that a 70–year-old in 2050, for example, will be healthier than a 70–year-old today, and will require less health care and old-age care. This assumption is quite controversial – for example, the DREAM Model does not make these assumptions either.

39. The simulated impact of these assumptions on the fiscal accounts are shown in Table 3. There are several important features of these results. First, total expenditures in 2050 drop by more than four percentage points – from 61.6 to 57.3 percent of GDP.10 Not surprisingly, most of this drop comes from a decrease in pension expenditures, which account for about 2½ percentage points of the total decline. The rest of the decline is due to an increase in GDP, which lowers most other expenditure categories when expressed as a percent of GDP. Second, although pension expenditures fall sharply, the overall effects of ageing are not completely offset. This is because, although longevity has been rising for several years now, the increase in age limits does not start until 2019, and they essentially “grandfather” a large number of retirees that retire before 2019. In addition, those individuals that retire between 2019 and 2027 (when the Welfare Agreement is fully effective) still have longer retirements than today’s retirees. Third, the estimated costs of ageing with respect to health care and old age care are still substantial – increasing 7 percentage points from 2008 to 2050. Finally, it is important to note that the effects of the Welfare Agreement and the Jobs Plan – while substantial – are not enough to attain fiscal sustainability. Indeed, as the table shows, the fiscal sustainability gap is about 2 percent of GDP (compared to about 4 percent of GDP without the policy changes).

Table 3.

Denmark: Effects of the Welfare Agreement on Fiscal Sustainability, 2008–2050

(Percent of GDP)

article image
Source: IMF staff calculations.

Other Policy Changes and Policy Proposals

40. More recently, the authorities have initiated a number of new policy changes, which will be phased in over the next few years. First, there was a cut personal income taxes, but this change was about offset by an indexation of energy taxes and a cancellation of a planned cut in the labor market contribution. Second, the continued tax freeze has led to a significant fall in some taxes, particularly property taxes. Third, new objectives to address energy and global warming challenges have also deteriorated long-term public finances (although these initiatives may yield some benefits in later years). Finally, there are announced initiatives to boost public consumption (including the recent wage hikes) and public investment. These initiatives are also hard to evaluate, since they have a negative effect on the fiscal balance in the short run, but could boost national income in the future, depending on the impact of public investment on the capital and the response of private investment. The authorities have estimated that the overall effect of these policy changes is an increase in the sustainability gap of about ¾ percent of GDP.

41. Finally, the Labor Commission recently released a number of proposals to improve fiscal sustainability and boost labor supply over the long run. The most important of these proposals (from a fiscal standpoint) is to bring forward the planned increases in age limits on pensions. Other proposals include a reduction in the duration of unemployment benefits from 4 years to 2 years, an increase the frequency of contact between the unemployed and the employment centers, a shift the focus of active labor market programs from education to on-the-job training, , and an improvement in the recruitment of foreign labor. Based on an alternative simulation, bringing forward the age limits has a substantial effect on fiscal sustainability – about 2 percentage points of GDP. In other words, if this proposal alone was implemented, the fiscal gap would be about ¾ of GDP. The other proposals made by the Labor Commission are more difficult to evaluate in this framework, but it is likely that they would close the gap, and this is indeed the view taken by the Labor Commission itself.

References

  • Barro, Robert J., and Xavier Sala-i-Martin, 2004, Economic Growth, 2nd Edition, Cambridge, MA: The MIT Press.

  • Blanchard, Olivier, Jean-Claude Chouraqui,Robert P. Hageman, and Nicola Sartor, 2004, “The Sustainability of Fiscal Policy: New Answers to an Old Question”, OECD Economic Studies, No. 15 (Autumn): 7 -36.

    • Search Google Scholar
    • Export Citation
  • Borjas, George J., 2003, “The Labor Demand Curve Is Downward Sloping: Reexamining the Impact of Immigration on the Labor Market,” Quarterly Journal of Economics, November: 1335 -74.

    • Search Google Scholar
    • Export Citation
  • European Commission, 2005, “The Economic Impact of Ageing Populations on the EU25 Member States,” Economic Papers #236.

  • International Monetary Fund, 2001, World Economic Outlook, Washington, DC.

  • Ottaviano, Gianmarco I.P., and Giovanni Peri, 2006, “Rethinking the Effects of Immigration on Wages,” NBER Working Paper #12497.

  • Schou, Poul, 2006, “Immigration, Integration, and Fiscal Sustainability,” Journal of Population Economics, 19:671-89.

1

Prepared by Allan D. Brunner (EUR).

2

Unfortunately, earnings data by ancestral categories were not available. Instead, it is assumed that, after adjusting for differences in experience and education, workers earn the same average wage regardless of ancestry.

3

Data on labor force participation rates and unemployment rates by age and ancestry are available from 1991 to 2006. There is some variation in labor market activity patterns across time. However, I focus on 2006, which represents that largest number of immigrants and descendants, as a share of the total population.

4

These data were compiled by the DREAM model (www.dreammodel.dk) based on microdata. Because of insufficient data for certain groups, descendants are assumed to have the same demands for public benefits as native Danes, and all individuals older than 70 years old are assumed to place the same demands on the public sector.

5

Note that if one is willing to make forecasts of the primary surplus out through infinity, then there is no need to set a debt target, since the discounted value of that target goes to zero as s approaches infinity.

6

Using a weighted sum over population segments implies that groups are perfectly substitutable. That is, after adjusting for differences in productivity, a 55–year-old individual with a basic education is perfectly substitutable for a 20–year-old individual with a Ph.D. Borjas (2003) and Ottaviano and Peri (2006) find evidence that this assumption is correct for U.S. workers. However, some preliminary work (not presented here) suggests that Danish workers are much more substitutable.

7

Since labor’s share of output is 0.55, the growth rate in labor efficiency is equivalent to a 1 percent rate of growth in total factor productivity (TFP).

8

A review of the literature shows a wide range of estimates for the average effective tax rate on capital for Denmark—from 15 to nearly 50 percent.

9

Indeed, one way to interpret the calculations in this version of the note is that tax increases will be levied on labor and that labor is perfectly immobile.

10

This estimate is somewhat less than the estimate provided by the Ministry of Finance in their 2007 Convergence Program. Indeed, the Ministry is more optimistic on almost all aspects of the calculation (discussed further above). In contrast, this estimate is much more in line with projections made by the DREAM Model and the Economic Council.

References

  • Andersen, Torben and Michael Svårer (2007), “Flexicurity—Labor Market Performance in Denmark”, CESifo Economic Studies, vol. 53, pp. 389 –429.

    • Search Google Scholar
    • Export Citation
  • Ball, Laurence (1997), “Disinflation and the NAIRU”, in Reducing Inflation: Motivation and Strategy edited by Christina Romer and David Romer, pp. 167 –94. Chicago, IL: University of Chicago Press for the NBER.

    • Search Google Scholar
    • Export Citation
  • Ball, Laurence and N. Gregory Mankiw (2002), “The NAIRU in Theory and Practice”, Journal of Economic Perspectives, vol. 16, pp. 115 -36.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier and Peter Diamond (1994), “Ranking, Unemployment Duration, and Wages”, Review of Economic Studies, vol. 61, pp. 417 –34.

    • Search Google Scholar
    • Export Citation
  • Blanchard, Olivier and Lawrence Katz (1999), “What We Know and Do Not Know About the Natural Rate of Unemployment”, Journal of Economic Perspectives, vol. 11, pp. 51 –72.

    • Search Google Scholar
    • Export Citation
  • Gordon, Robert J. (1998), “Foundations of the Goldilocks Economy: Supply–Shocks and the Time Varying NAIRU”, Brookings Papers on Economic Activity, 2, pp. 297- 333.

    • Search Google Scholar
    • Export Citation
  • Jensen, P., Michael Rosholm, and Michael Svårer (2004), “The Response of Youth Unemployment to Benefits, Incentives, and Sanctions”, European Journal of Political Economy, vol. 19, pp. 301– 16.

    • Search Google Scholar
    • Export Citation
  • King, Robert, James Stock, and Mark Watson (1995), “Temporal Instability of the Unemployment-Inflation Relationship”, Economic Perspectives of the Federal Reserve Bank of Chicago, pp. 2- 11.

    • Search Google Scholar
    • Export Citation
  • Perry, George (1970), “Changing Labor Markets and Inflation”, Brookings Papers on Economic Activity, 3, pp. 411 –48.

  • Rosholm, Michael (2001), “Cyclical Variations in Unemployment Duration”, Journal of Population Economics, vol. 14, pp. 173– 91.

  • Rosholm, Michael and Michael Svårer (2008), “Estimating the Threat Effect of Active Labor Market Programmes”, Scandinavian Journal of Economics, vol. 110, pp. 385 –401.

    • Search Google Scholar
    • Export Citation
  • Svårer, Michael (2007), “The Effect of Sanctions on the Job Finding Rate: Evidence from Denmark”, IZA Discussion Paper 3015.

  • Staiger, Douglas, James Stock, and Mark Watson (1997), “How Precise are Estimates of the Natural Rate of Unemployment?”, in Reducing Inflation: Motivation and Strategy edited by Christina Romer and David Romer, pp. 195 –242. Chicago, IL: University of Chicago Press for the NBER.

    • Search Google Scholar
    • Export Citation
  • Staiger, Douglas, (2001), “Prices, Wages, and the U.S. NAIRU in the 1990s” in The Roaring Nineties edited by Robert Solow and Alan Krueger, pp. 3 –60. New York, N.Y.: Russell Sage Foundation/The Century Fund.

    • Search Google Scholar
    • Export Citation
  • Stock, James and Mark Watson (1999), “Forecasting Inflation”, Journal of Monetary Economics, vol. 44, pp. 293 –335.

1

Prepared by Jay Surti (EUR).

2

A similar phenomenon of trend reversal in unemployment and a waning trade–off with inflation was noted in the U.S. during the 1990s, and instigated several studies of the robustness of the Phillips curve relation, the evolution of the NAIRU, and reasons for its apparent decline. Ball and Mankiw (2002) and Blanchard and Katz (1999) provide partial surveys of both, the underlying issues and the literature. King and others (1995) carried out an early empirical reassessment, followed by the studies of Gordon (1998) and Staiger and others (1997, 2001).

3

For a comprehensive overview of the various reforms implemented in the last 15 years, see for e.g., Andersen and Svårer (2007).

4

A theoretical treatment of how unemployment duration may affect the hazard rate out of unemployment into employment—negative duration dependence—is contained in Blanchard and Diamond (1989); Blanchard and Katz (1999) reflect on macroeconomic implications of negative duration dependence, and consequently, of labor market reforms for the NAIRU. For Denmark, in a study of the labor market in the 1980s, Rosholm (2001) concluded available evidence as providing no support for the hypothesis of negative duration dependence, neither at long durations, nor when aggregate unemployment is high.

6

Education costs in Denmark are not high as is the case in the U.S., otherwise paying off student loans would provide an additional incentive to stay employed, lowering their reservation wage.

7

Ball (1997) applies Perry–weighting to assess the impact of unfavorably changing demographics on rising trend unemployment in the OECD countries during the 1980s.

8

See Chapter I of this Selected Issues Paper.

9

Most data series are taken from the OECD’s analytical database and are at quarterly frequency. The sole exception is data on the age composition of the labor force which was obtained from Statistics Denmark.

10

The estimated cointegrating vector being (1, –0.8).

11

Staiger and others (2001) discuss of the underlying rationale.

12

The methodology is quite standard, and entails simultaneous estimation of the NAIRU and the Phillips curve via a Kalman smoother applied to a state–space model. See e.g., Gordon (1998), and Staiger and others (1997, 2001) for details.

13

Alternative specifications of the NAIRU process; for e.g., an AR(1) specification did not yield materially different results.

Denmark: Selected Issues
Author: International Monetary Fund
  • View in gallery

    Denmark: Ratio of Working–Age to Elderly Population, 2008–2050

  • View in gallery

    Denmark: Population Distribution, 2008, 2025, and 2050

    (Thousands)

  • View in gallery

    Denmark: Composition of Population, by Ancestry, 1981–2050

    (Millions)

  • View in gallery

    Denmark: Composition of Working-Age Population, by Ancestry, 2008-2050

    (Millions)

  • View in gallery

    Denmark: Composition of Elderly Population, by Ancestry, 2008-2050

    (Millions)

  • View in gallery

    Denmark: Labor Force Participation Rates, by Age and Ancestry, 2006

  • View in gallery

    Denmark: Unemployment Rates, by Age and Ancestry, 2006

  • View in gallery

    Denmark: Public Consumption for Men, by Age, 2005

    (Thousands of DKR)

  • View in gallery

    Denmark: Public Transfers for Men, by Age, 2005

    (Thousands of DKR)

  • View in gallery

    Denmark: Various Measures of the Labor Force, 2008–2050

    (Percent change)

  • View in gallery

    Denmark: Real GDP and Effective Labor Force, 2008–2050

    (Percent change)

  • View in gallery

    Denmark: The Effects of the Welfare Agreement on the Old-Age Dependency Ratio

    (Mandatory pensioners as a percent of working-age population)