Extensive literature estimates the impact of taxes on labor, but much less is known about how taxes affect the supply of capital. In theory, wealth taxes may affect saving and consumption decisions, business creation, tax compliance, and inter-generational wealth transmission. Capturing the precise economic effects of taxing wealth poses formidable empirical challenges, however, due to the lack of reliable micro data on household wealth, particularly amongst the wealthiest.
This paper provides new evidence from Denmark, which used to impose one of the world’s highest marginal tax rates on wealth, and therefore has a rich and comprehensive annual dataset of taxable wealth. In addition, variations in the marginal tax rate—leading up to the eventual abolition of the tax in 1997—allow for an estimation of the causal effects of taxing wealth.
At the aggregate level, it does not appear that the abolition of the wealth tax led to runaway inequality. Wealth inequality is markedly lower in Denmark than in the US, for example, with the top 1% richest households owning around 20% of total wealth in Denmark in 2012, compared to almost 40% in the United States. The gap between the two countries has actually widened over time; from the late 1990s wealth inequality stabilized in Demark—despite the ramping down of the wealth tax—while it continued to increase in the United States. This does not necessarily imply that wealth taxes had no effect on wealth accumulation. It simply means that if the wealth tax cuts caused wealth to grow faster at the top of the distribution then this effect must have been offset by factors working in the opposite direction.
This paper used a difference-in-differences approach to estimate the results of a reform that had a differential effect on singles and married couples within a certain range of the household wealth distribution. The authors find that although the two groups had parallel wealth trends in the five years before the reform, the wealth level of the couples—who had become exempt from wealth taxation—grew more quickly in its aftermath.
What matters most for tax design, however, are the long-run effects of taxes. The paper also constructs a lifecycle savings model, which indicates that the effect of the wealth tax reform on the stock of wealth grows for about 25 years and then stabilizes. After 30 years, wealth is 30% higher than it would have been without the reform. The long-run elasticity of wealth with respect to the net-of-tax rate of return is around 0.7, with the bulk of the effect coming from behavioral changes due to the increased incentive to save, rather than from mechanically faster wealth growth rates as a result of the higher net-of-tax rate of return. Wealth growth effects are larger for the very wealthy than for the moderately wealthy, both in the short and in the long run.
The idea of taxing top-end wealth has attracted considerable recent interest, in the United States and globally. Elasticity parameters such as those estimated here are vital for policy makers looking to assess both the equity and efficiency impacts of these wealth taxes.
Empirical difficulties have limited the available evidence on the effect of taxes on the supply of capital, but this new research exploits the gradual ramping down of Denmark’s wealth tax to estimate the effects of taxing wealth. At the aggregate level, inequality stagnated in the aftermath of the tax’s abolition, but difference-in-differences estimates indicate that the groups who benefitted from the tax rate cuts did experience faster rates of wealth growth. Overall, the long-run elasticity of wealth with respect to the net-of-tax rate of return is estimated at around 0.7. Elasticity parameters such as this are vital for policy makers looking to assess both the equity and efficiency impacts of wealth taxes.
What are the economic effects of taxing household wealth? While an enormous literature estimates the impact of taxes on labor supply and taxable income, much less is known about how taxes affect the supply of capital. The lack of evidence makes it hard to assess the desirability of taxing top-end wealth, a proposal that has gained interest in the United States and globally.
In theory, wealth taxes may affect saving and consumption decisions, portfolio allocation, business creation, tax avoidance and evasion, and the transmission of wealth across generations. But providing compelling evidence on the magnitude of these behavioral responses has proven difficult.
This article reports on the findings from our recent research on these questions. Our laboratory is Denmark, which used to impose one of the world’s highest marginal tax rates on wealth. This tax was reduced starting in 1989 and abolished in 1997. This policy experiment is useful for learning about behavioral responses among the wealthiest segments of the population.
Do aggregate wealth trends provide a smoking gun for behavioral responses?
To start with, it is useful to document the evolution of wealth inequality in Denmark. Figure 1 plots the evolution of the top 1% wealth share in Denmark and the U.S. from 1980 to 2012. Several insights are worth noting. First, wealth inequality is markedly lower in Denmark than in the U.S. In 2012, the top 1% richest households owned about 20% of total wealth in Denmark, whereas they owned almost 40% of total wealth in the U.S. Average wealth in the population was similar in the two countries, but households in the top 1% were twice as wealthy in the U.S as in Denmark. Second, the gap between the two countries has widened over time. Top wealth shares were increasing in both countries until the late 1990s, but then they diverged as wealth inequality continued to increase in the U.S but stabilized in Denmark, despite the latter country’s reduction and ultimate abolition of the wealth tax.
These aggregate trends do not provide a “smoking gun” for any impact of wealth taxes on wealth accumulation, but this does not necessarily imply that wealth taxes had no effect on wealth accumulation. It simply means that if the wealth tax cuts caused wealth to grow faster at the top, this must have been offset by factors working in the opposite direction. One such factor is pension wealth; in Denmark, pension wealth rose significantly in the 1990s and 2000s, from around 50% of national income in the late 1980s to 178% in 2014, in part due to the automatic enrollment of individuals in employer-sponsored pension plans. Because pension wealth is relatively equally distributed, rising pension wealth tends to reduce inequality.
The challenge of estimating the effects of wealth taxes
Capturing the economic effects of taxing wealth poses formidable empirical challenges. First, to say anything about household wealth accumulation, we need reliable data on wealth. Few countries collect administrative micro data on household wealth. Second, to understand how wealth taxes affect wealth accumulation, we need variation in wealth taxation that allows for the estimation of causal effects. Few countries have wealth taxes today, though a dozen or so OECD countries did have wealth taxes up until the 1990s. Third, because wealth is relatively concentrated, it is crucial to estimate behavioral responses for the wealthiest individuals, whose data are often particularly difficult to access.
Denmark as a laboratory
Denmark offers the data and tax variation that allow us to overcome the challenges described above. Denmark taxed wealth on an annual basis until 1997, which meant that it has a rich and comprehensive dataset on taxable wealth (i.e., the total net wealth of households, excluding pension wealth). The Danish wealth tax also changed through time; until a major reform in 1989, wealth was taxed at a flat rate of 2.2% above a threshold located around the 98th percentile of the household wealth distribution. The 1989 reform reduced the marginal wealth tax rate from 2.2% to 1%, however, and doubled the exemption threshold for married couples (but not for single individuals).
Effects of wealth taxes
One implication of the 1989 reform is that married couples in a certain range of the household wealth distribution became exempt from wealth taxation, while singles in that same range experienced a smaller reduction in the wealth tax rate. This variation allows us to compare wealth accumulation in these two groups over time using difference-in-differences estimation techniques.
Figure 2 plots the difference in the logarithm of taxable wealth between couples and singles in the affected range over time. The difference in each year is measured relative to the pre-reform year, 1988. Two key insights emerge from the figure. First, there is hardly any difference in the wealth trends of the two groups prior to the reform. While there were some differences in the early 1980s, the trends are virtually parallel in the five years leading up to the reform. Second, the wealth levels of the two groups start diverging immediately after the reform as couples in the exempted range accumulate taxable wealth at a faster rate than singles. The difference in wealth grows over time, consistent with an increased savings rate of exempt couples relative to non-exempt singles.
Overall, this figure provides evidence of behavioral responses to the reduction in wealth taxation. The effect on the log of taxable wealth is equal to 0.186 in the last post-reform year (1996), i.e. an increase in average wealth of about 18% over 8 years.
From the short to the long run
The data allow us to estimate the effects of the 1989 wealth tax reform in the short to medium run. What matters most for tax design, however, are the long-run effects of taxes. This is particularly true for wealth taxes due to the dynamic and slow-moving nature of wealth accumulation, which mean that it takes a long time for the effects to play out.
To get at the long-run effects of the Danish wealth tax, we construct and calibrate a lifecycle savings model with utility for end-of-life wealth (due to bequest or other utility-of-wealth motives). The parameters of the model are calibrated to be consistent with the difference-in-differences estimates described above. The model allows us to simulate the long-run effects of wealth tax cuts on wealth accumulation at the top-end of the wealth distribution.
Simulation of long-run effects
Figure 3 shows the simulated effects over 30 years. The blue series shows the estimated effects of the 1989 reform over the short run (using the same data and population as in Figure 2), while the orange series shows the simulated effects (which match the empirical short-run effects reasonably well) over the long run.
The figure shows that the effect of the wealth tax reform on the stock of wealth grows for about 25 years and then stabilizes. After 30 years, wealth is 30% higher than it would have been absent the reform. Though this effect is large, the underlying tax incentive for saving is also large. Therefore, the long-run elasticity of wealth with respect to the net-of-tax rate of return is still reasonable, around 0.7. Part of this effect is purely mechanical: for a given saving rate out of pre-tax income, a larger net-of-tax rate of return mechanically increases wealth over time. But the bulk of the effect we estimate is behavioral: with a larger net-of-tax rate of return, the savings rate increases.
What about the very wealthy?
The results presented above capture responses by moderately wealthy households, those in the 98-99th percentile of the wealth distribution. In another research design, we exploit the fact that, among the very wealthiest households, some face a zero marginal tax rate on wealth due to a tax ceiling that limits the total average tax rate from all personal taxes (income, social security, and wealth taxes). As a result, the tax reform had differential impacts on those bound and unbound by the ceiling. We find larger effects for the very wealthy than for the moderately wealthy, both in the short and in the long run.
What have we learned?
We have made progress towards addressing one of the most important questions in public finance: What is the long-run effect of capital taxation on the supply of capital? The answer to this question is critical for assessing the degree to which governments should tax capital income or wealth. The empirical challenges discussed in the beginning have prevented serious progress on answering this question, but the Danish setting allows us to estimate some of the key elasticities. Based on a quasi-experimental analysis, we find clear reduced-form effects of wealth taxes on taxable wealth in the short to medium run, with larger effects on the very wealthy than on the moderately wealthy. Based on a calibration analysis (that respects the quasi-experimental results), we find that the long-run elasticity of taxable wealth with respect to the net-of-tax return is sizeable at the top of the distribution. The elasticity parameters we estimate would help policy makers assess the trade-off between efficiency and equity when setting wealth taxes as well as other forms of capital taxes.
This article summarizes ‘Wealth taxation and wealth accumulation: Theory and evidence from Denmark’ by Katrine Jakobsen, Kristian Jakobsen, Henrik Kleven, and Gabriel Zucman, published The Quarterly Journal of Economics in February 2020.
Katrine Jakobsen is at the University of Copenhagen. Kristian Jakobsen is at Den Sociale Kapitalfond. Henrik Kleven is at Princeton University. Gabriel Zucman is at the University of California, Berkeley.