Even Denmark’s top-notch pension system can be improved. The backbone of the system—that you work one extra year for every extra year you’re expected to live—means future generations will spend a considerably smaller share of their adult lives in retirement than current generations. This is not inter-generationally fair. Also, keeping pension contribution rates the same throughout your working life helps make sure you save enough for retirement. But a life-long constant contribution rate can be too rigid, and ironically, it might mean saving too much as retirement ages go up. I’ll discuss these issues and more. I’ll also suggest some possible ways forward.
In my last post, I argued that Denmark has a strong retirement system (link)—one that might also inspire other countries (link).
This system has contributed to sound macroeconomic balances. Danish public finances are strong, with very low public debt both now and looking ahead. This is thanks in no small part to the indexation of the retirement age to life expectancy.
Denmark is also a net international lender, with large surpluses on its current account, driven largely by high private savings (not least in pensions). The system also ensures an adequate income in retirement for most Danish pensioners, resulting in high coverage ratios. Consequently, the incidence of poverty among Danish retirees is low. Finally, pension funds are large enough to benefit from economies of scale—a goal that other countries aspire to achieve (link). Overall, it is a pension system that many countries envy.
However, this does not mean there is no room for improvement. Even Denmark’s system can be refined. In this post, I will discuss areas for improvement and suggest possible solutions.
A cautionary note at the outset: most of the strengths of the Danish pension system I highlighted in my previous post are largely undisputed and are widely acknowledged by pension economists, the pension industry, and policymakers. That may not be the case for the considerations I explore here. Therefore, before diving in, I want to make it clear that these are my own views.
Let’s get started.
Consideration 1: The system is not fair across generations
First, as I mentioned in my previous post (link), the 2006 Welfare Agreement established a connection between the retirement age and life expectancy, with a target of 14.5 years in retirement. The goal is for the link between life expectancy and retirement age to be 1:1—for every additional year Danes are expected to live, they would also work one more year. However, since Danes are currently spending more than 20 years in retirement on average, the retirement age is now being increased at a pace that exceeds the rise in life expectancy, to reduce the average retirement period toward the 14.5-year target. At the same time, increases to the retirement age are capped at a maximum of one year every five years.
Simple mathematics tells us that a 1:1 indexation of the retirement age to life expectancy means that future generations will spend a smaller proportion of their adult lives in retirement than current retirees (see Figure 1): you live longer, so you work longer, but you do not stay longer in retirement, hence you will spend a smaller proportion of your adult life in retirement.
Figure 1. Proportion of adult life spent in retirement for different generations. ‘G1900’ means the generation born in 1900, for example. Source: Pensionskommissionen (link).
Danes born in the 1950s—those now in retirement—will spend almost a third of their adult lives in retirement. By contrast, children born today can expect to spend only about a fifth of their adult lives in retirement.
Is it fair, from an inter-generational perspective, that current retirees enjoy so much more time in retirement than future retirees? I think not, but the solution is not straightforward.
This very question—whether the 1:1 indexation scheme is too “strict”—is at the centre of the Danish pension debate now.
One alternative could be a 0.8:1 indexation; that is, for every additional year of life expectancy, the retirement age would rise by 0.8 years (or 9.6 months), but other types of indexations are also possible. However, loosening the 1:1 indexation would come at a (high) cost to public finances, as I showed in my previous post. A well-balanced compromise is therefore needed.
Consideration 2: Contributions could be made more flexible, particularly for younger people
Currently, the typical second-pillar occupational pension scheme requires Danes to contribute a fixed proportion of their salary to pension savings—often 10%, 12%, 15%, or even more—from their first day of work until retirement. There are, however, arguments for making the system more flexible, especially in the early stages of working life, though potentially also later.
In the early phase, people often have relatively low incomes but face substantial expenditures, such as raising children or buying their first home. Yet, the current system requires you to pay the same contribution rate as later in life. Here, some flexibility might be helpful. Personally, I would favour a system where individuals could pause or reduce their pension contributions for a period—say, a year or two—when having a child or purchasing a home.
This should, in my view, be a voluntary opt-out provision, with clear information about the impact on future pension income. Pausing pension contributions during periods of high expenditure would effectively increase disposable income in those years, thereby also lowering borrowing needs, see also my next ‘consideration.’
In the later phase—specifically, in the years before retirement—people can currently start drawing their pension up to three years before the official retirement age. However, given that future generations may have to work until they are 77, is a three-year window sufficient? I do not hold strong views on this point, but it is worth considering.
An alternative to the system I propose would be to start with lower contribution rates for younger workers, increasing them as they get older. Research shows that a flat contribution rate is rarely optimal in mandatory pension plans, while a rising rate often is (link, link). For example, in Switzerland, pension contributions are rising with age. However, I personally favour a system that allows for voluntary opt-outs when young, rather than one that mandates the same rates for everyone at all stages in life, whether that rate is flat or increasing.
Consideration 3: Contribution flexibility for young people could reduce Danish households’ high leverage
There is another important reason to allow for pension contribution flexibility among young people: it could help to reduce Danish households’ otherwise high levels of debt.
Although Danes accumulate significant pension savings, they also carry substantial debt. Indeed, Danish households are among the most indebted in the world, cf. Figure 2. There is reason to believe—and research shows (link)—that high mandatory pension savings contribute to household indebtedness.
Figure 2. Household debt as a percentage of household net disposable income, different countries, 2023. Source: OECD and J. Rangvid.
This connection raises two concerns. First, there is a financial stability concern. Danish households have large balance sheets, with both substantial savings and substantial debt. However, pension savings are long-term and illiquid, and making them liquid comes at a high cost. In the event of financial hardship—such as an economic crisis—the high debt burden could become problematic. That is, you might have a lot of pension savings, but they do not really help you when the crisis hits, because they are illiquid.
Second, Danish households may end up paying higher fees to financial intermediaries (pension funds and banks) than they would otherwise have done. If households are compelled to take on debt to fund pension contributions—which is in fact what happens when you are forced to pay a substantial fraction of the your income into your pension scheme even though you have large borrowing needs—they may face higher financial costs than they would with more flexibility, a hidden cost for the household.
This balance sheet expansion is particularly challenging for younger people. In a paper with Hebsgaard and Ramlau-Hansen (link), we calculated the leverage of Danish households—i.e. their assets and liabilities relative to their equity (assets minus liabilities)—over the life cycle. This is shown in Figure 3.
Figure 3. The figure shows the average Danish household’s bank loans, mortgage loans, value of real estate, pension savings, and stock and bonds, and bank deposits, all relative to the household’s equity, over the life cycle. Source: Hebsgaard, Ramlau-Hansen, and Rangvid (link).
Young Danes are highly leveraged. They hold bank and mortgage debt amounting to three times their equity. If they could temporarily opt out of (or reduce) their pension contributions, they might not need to borrow so heavily to finance their first home or to cover the costs of raising children. This again supports the case for introducing flexible contribution clauses—allowing young people to pause pension payments when they have high financing needs.
Consideration 4: Many Danes risk saving too much if the system is not adjusted
Already today, most Danish retirees enjoy adequate income in retirement. In my previous analysis (link), I showed that the vast majority of today’s retirees have coverage ratios above 70%—a very positive outcome.
However, looking to the future, and looking at the ‘all else equal’ situation, Danes will work longer (due to the 1:1 indexation of retirement age to life expectancy) and spend fewer years in retirement. Just a few years ago, the retirement age was 65; by 2100, it is projected to reach 77—a difference of 12 years.
Moreover, current retirees can expect to spend around 20 years in retirement. In the future, the target is 14.5 years. A reduction of more than 5 years.
Taken together, this means many Danes risk saving more than necessary for their retirement. It is likely that future retirees’ income from private pensions could be up to 50% higher than that of today’s retirees.
Table 1 provides a (very) simple example to illustrate this point.
Table 1. A simple example showing how savings and income at retirement vary by retirement age and retirement length.
Imagine that a 25-year-old Dane starts saving 100 DKK per year until retirement at age 65—that is, saving for 40 years. By retirement, this person would have accumulated 4,000 DKK in pension savings, the first row in the table.
Now imagine that this pension is intended to cover 20 years in retirement. Ignoring investment returns and discounting for simplicity (effectively similar to a salary-adjusted real interest rate of zero), this would generate an annual retirement income of 200 DKK.
This scenario—retiring at 65 with 20 years of expected retirement—closely resembles the situation just a few years ago (see Figure 1 in my previous post).
However, if the same 25-year-old now must work until age 70 and only spends 15 years in retirement (as implied by the current system’s 1:1 indexation of retirement age to life expectancy), their pension savings at retirement would be around 4,500 DKK—about 13% higher than in the earlier scenario. Yet, this pot would now be spread across only 15 years of retirement, meaning an annual pension income of 300 DKK—50% higher than before.
Of course, this is a very simplified example. Nevertheless, more precise and realistic calculations that incorporate the full details of the system yield similar results, indicating that private pension income could increase by around 50% in the future.
It is important to note that this calculation applies only to private pension savings. Since the state pension is not affected by the increase in life expectancy, total pension income—comprising both the state pension and private pensions—will not rise by 50%, but perhaps by around 20% on average.
Yet, if most Danes are already saving enough, why would they need significantly higher incomes from their private pensions? There is fiscal space to consider reducing the strict 1:1 indexation and introducing more flexibility in pension contributions.
Consideration 5: Without reform, some may choose to retire early
If no adjustments are made—meaning retirement ages continue to rise and the proportion of an adult’s life spent in retirement falls—there is a significant risk that a substantial fraction of future retirees will accumulate enough pension wealth to leave the labour market earlier than the official retirement age would otherwise mandate.
This would undermine the very aim of linking retirement age to life expectancy. If people save so much that they can afford to retire early, they will stop paying income taxes earlier too (if they retire early). The strong public finances would not be so strong anymore.
A counterargument is that, so far, people do tend to work longer as the retirement age rises. This can be seen in Figure 4, which shows how employment rates across different cohorts change when these cohorts become eligible for the state pension.
Figure 4. Employment rates of different cohorts across their different retirement years. Source: Pensionskommissionen (link).
Take the March 1953 cohort, for example. This group became eligible for state pension at age 65. As shown in Figure 2, when they reached that age, their employment rate dropped sharply. In other words, these people retired when they became eligible for state pension.
The same sharp drop was not observed for the March 1954 or September 1954 cohorts, who could only retire at ages 65.5 and 66, respectively.
However, when the March 1954 cohort reached their retirement age, their employment rate fell as they withdrew from the labour market. The September 1954 cohort did not initially withdraw, but when they turned 66 and became eligible, many also left the labour market.
This means that, so far, the evidence suggests that people do indeed work longer when the state pension age rises. But will this continue to hold true as people face working even longer, and accumulate sufficient pension savings to retire earlier?
Consideration 6: ATP needs fundamental reform
ATP (link) remains the largest pension fund in Denmark, and all working Danes are required to pay into it. For those who also participate in other labour-market pension schemes, ATP’s role is relatively minor. However, ATP is the only private pension saving for those without a labour-market pension—typically low-income earners.
For several years, I have (together with my colleague Henrik Ramlau-Hansen) argued that ATP is an ill-designed pension product. The main issue is that ATP is structured to guarantee a nominal pension. One might argue that a nominal guarantee is valuable in the payout phase. But there is little reason to believe that an 18-year-old entering the labour market today needs to know with certainty what their nominal pension income will be in 40, 50, or even 60 years. For these younger workers, it would be far more important to aim for the highest possible pension income in real terms, even if that means accepting some uncertainty in nominal outcomes.
The problem is that ATP needs to invest conservatively to guarantee future nominal payouts. But conservative investments yield low returns. As a result, Danish workers are forced into a pension fund that invests too heavily in safe assets—leading to unnecessarily low pension incomes. Ramlau-Hansen and I have proposed an alternative model for ATP that maintains a similar risk level but could deliver pension incomes that are 15-35% higher, depending on different assumptions.
While the bulk of ATP’s investments are conservative, a small portion is invested in high-risk assets—arguably too high-risk. These investments are heavily leveraged, concentrated in Danish and small-cap equities, and include substantial allocations to illiquid assets. If this sounds like a hedge fund, you’re not wrong.
Unfortunately, the performance of this high-risk portfolio has been disappointing: average returns have been -15% per year for the past three years. By contrast, other high-risk Danish pension funds have generated returns of +5% per year.
In earlier periods, ATP performed better, but the pattern of high returns early on followed by very poor returns later has not paid off. Over the long run, ATP’s returns have roughly matched global stock market returns—no better, no worse—but with much higher risk.
And that’s not all: ATP has underperformed other pension funds in terms of returns to its members’ saving (i.e., the guaranteed portion), and its payouts cannot keep up with inflation, meaning they will fall in real terms. There are also concerns about corporate governance at ATP, and ATP is not subject to the same regulations as other pension funds (e.g. ATP is not required to meet solvency requirements).
I could go on, but this post is not meant to be a comprehensive critique of ATP—it’s about challenges in the Danish pension system as a whole. Nevertheless, ATP’s issues are significant for the wider system, particularly because ATP is important for low-income Danes.
Conclusion
Denmark has a very strong pension system—one that many countries envy. It secures macroeconomic balance and provides adequate retirement income for most retirees. However, no system is perfect. In this blog post, I have described ways to improve an already solid pension system and ensure it remains fit for the future.
