Best quality data
The slow pace at which wealth distribution typically changes, combined with the challenges posed by assembling wealth distribution data, makes it difficult to determine inequality trends over relatively short periods, apart from times of war, revolution or market collapse. It is therefore prudent to look first at longer-run trends, and also to the United States, since it is the country which contributes most to global household wealth, produces the best quality data on personal assets, and has estimates of wealth inequality going back a century.
Figure 1 shows the share of wealth owned by the top percentile and decile (i.e. the top 1% and top 10%) of wealth holders from 1910 onwards. The corresponding shares for income recipients are also displayed for reference and show a more pronounced pattern. The share of the top income percentile peaked at 20% in 1928 and then eased down below 10% in the 1950s, to bottom out at 8% in the 1970s. It then rose sharply after 1980, eventually returning to the peak level of the late 1920s. The share of the top income decile followed a similar pattern, although the decline in the early 1940s was more abrupt, and the recent rise has been more sustained.
Wealth inequality is much higher than income inequality. This is a worldwide phenomenon. In Figure 1 the share of wealth for the top percentile is typically around 20 percentage points higher than share of income, and for the top decile it is typically 30 percentage points higher. But the curves for income and wealth are broadly similar in shape, trending downward from the late 1920s to the 1970s, and then rising. However, in contrast to the income inequality series, the share of the top wealth percentile has shown little upward movement for the past two decades: from 1989 until the present day it has remained around 33%. On the other hand, the share of the top 10% ‒ which better reflects the broader trend – rose gently from 67% in 1989 to 72% in 2007 and then jumped to 74.5% in 2010. Figures just released for 2013 suggest a further rise to 75.3%. These findings point to an upward shift in wealth inequality in recent years.
While top decile wealth shares above 70% are high by international standards, they understate the true level of wealth inequality in the United States since the estimates are derived from the Survey of Consumer Finance (SCF), which explicitly omits coverage of a small number of holdings (roughly those of billionaires) at the apex of the pyramid. Our estimates make adjustments for these missing observations and consequently tend to produce slightly higher values for the share of the top decile and top percentile.
Interestingly, for the United States we find no significant trend in the top wealth shares in either direction. Our calculations suggest that the share of the top percentile was 38.5% in 2000, rose to 38.9% in 2007 and 2008, and then drifted down to 38.4% by mid-2014. These findings are broadly in line with the SCF evidence, although our estimate of the share of the top percentile is significantly higher than the SCF figure, reflecting the fact that we take account of the billionaires. For the top decile, we estimate that its share edged up from 74.6% in the year 2000 to 74.8% in 2007 and 2008, before dropping back to 74.6% again by 2013. Again, our figures broadly conform to the SCF data except for the upward move since 2007, which we do not observe. One possibility is that the SCF is improving its coverage of UHNW holdings, so that our top tail adjustment is becoming redundant – at least as far as the US figures are concerned.
Long-term trends in other countries
The United States is one of ten countries for which long-term wealth inequality series can be constructed. This group is conveniently subdivided into three Anglo-Saxon countries (Australia, the United Kingdom and the United States), four Nordic countries (Denmark, Finland, Norway and Sweden) and three continental European countries (France, the Netherlands and Switzerland). While data from different sources must often be spliced together to obtain these series, the overall conclusion is fairly consistent. In most cases, the share of the top 1% of wealth holders trended down from the 1920s to the 1970s, flattened out, and then edged up slowly. Over the whole period 1914-2010 the top percentile share fell in all countries except Switzerland, where no trend is evident. On the other hand, in most countries the wealth share of the next 4% (i.e. percentiles 95 to 99) in 2010 was close to its value in 1914.
Wealth inequality around the world
Although direct data on household wealth distribution are available for countries which collectively cover two thirds of the world’s population and about 90% of household wealth, in most cases the data are not of the same quality achieved by the SCF in the United States. Many countries have no direct wealth distribution data at all. The procedures that we have developed – including making use of rich list information on the top tails of wealth distribution – provide an alternative way of estimating wealth distribution and wealth inequality, and may be the only means of constructing plausible estimates for most countries in the foreseeable future. However, it should be borne in mind that the estimates are subject to a higher margin of error than usual.
Figure 1 provides a convenient benchmark for judging the levels of wealth inequality seen around the world. We use the term “medium inequality” to refer to a top decile share of 50-60%, a level reminiscent of Western Europe around 1980. Most developed countries fall in this category (see Table 1), so little change in inequality over time is evident here. The remaining developed economies typically have a top decile share of between 60% and 70% and are classed as “high inequality” – the level prevailing in the United States in the mid-20th century. The top decile share in Switzerland and the United States is above 70%: we label this as “very high inequality”, similar to that experienced in the United States a century ago. Nowadays, a share below 50% for the top wealth decile is uncommon. Belgium and Japan just squeeze below this “low inequality” threshold.
For emerging market economies, the classification system appears to shift upwards by a grade or more. The majority of countries, including many big players on the international scene – Brazil, India, Indonesia, Russia, South Africa and Turkey – qualify as “very high inequality”. According to our estimates, inequality in Russia is so far above the others that it deserves to be placed in a separate category. The remaining emerging market nations – including Chile, China, Korea and Taiwan – are classed as “high inequality”, except for Singapore and the United Arab Emirates, which rate as “medium inequality”. Interestingly, Singapore has the highest wealth per adult among emerging markets, and the United Arab Emirates is not far behind. This hints at the possibility that wealth inequality may tend to decrease as economies mature and average wealth increases, echoing the famous Kuznets’ hypothesis that income inequality typically increases during the early stages of development and later declines.
Other measures of wealth inequality
The arrangement in Table 1 changes slightly if countries are grouped instead according to the share held by the top percentile of wealth holders, with cut-off points of 40%, 30% and 20%. Switzerland and the United States now rate as “high inequality”, since the top percentile share is less than 40%. Similarly, Austria, Denmark, Germany and Norway register top percentile values below 30%, and hence qualify as “medium inequality”. However, five of these six countries are very close to the borderline – the exception being Switzerland for which the top percentile share is just 30.9%. Chile is another borderline case. The top decile share of 69.4% is just within the “high inequality” band, but the top percentile share of 41.8% suggests that Chile may be better viewed as “very high inequality”.
On balance, we favor using the top decile share as the primary indicator of inequality, because it is more broadly based and correlates well with the value of the Gini coefficient. In the context of wealth inequality, however, we recommend against using measures based on the share of the bottom half of the distribution, or the share of the lowest decile, since these values are highly sensitive to the large negative wealth holdings which are now increasingly common in countries with easy access to credit and high levels of student loans. In Denmark, for example, average wealth is negative for each of the three lowest deciles, although Denmark is not normally regarded as a particularly unequal nation.
Regional and global trends in wealth inequality
This year, for the first time, we have been able to construct consistent series for the distribution of wealth in all countries since the year 2000. The resulting data enable us to assess the direction and magnitude of trends in global wealth inequality. Our research suggests that countries often experienced a structural break in inequality trends around the time of the financial crisis. Prior to 2007 most countries show little change in inequality, or a slight decline; after 2007, wealth inequality has tended to increase.
Considering the entire period since 2000, Figure 2 shows that wealth inequality has increased in Latin America and Africa, and to a greater extent in India and China, but fallen slightly in Europe and North America, and also by a fraction in the world as a whole. For the Asia-Pacific region, the evidence is ambiguous: the share of the top wealth decile declined a little, but the share of the top percentile rose.
Splitting the period reveals markedly divergent trends before and after the financial crisis. From 2000 to 2007, inequality fell in every region except China and India. Since 2007, the share of both the top decile and the top percentile has risen in every region except North America. The reduction in wealth inequality during the early period was especially pronounced in Asia-Pacific, Europe and Latin America. The subsequent rises are more consistent across regions, North America excepted.
Inequality trends for individual countries
Inequality trends for individual countries are explored in more detail in Table 2, with countries listed in order of the increase in inequality since 2000. The most striking feature is the contrast in experience before and after the financial crisis. In the period from 2000 to 2007, 12 countries saw a rise in inequality while 34 recorded a reduction. Between 2007 and 2014, the overall pattern reversed: wealth inequality rose in 35 countries and fell in only 11. The reason for this abrupt change is not well understood, but it is likely to be linked to the downward trend in the share of financial assets in the early years of this century, and the strong recovery in financial assets since 2007.
To aid discussion, we use the term “slight rise” to denote an increase of 0.1-0.2 percentage points per year in the wealth share of the top decile. The corresponding ranges for a “rise” and a “rapid rise” are 0.2-0.5 and 0.5+ respectively. Similar labels are applied to decreases in wealth shares, while changes averaging between -0.1 and +0.1 percentage points per year are described as “flat”.
Over the entire period since 2000, nine countries have experienced a rapid rise in inequality, but only two had a rapid fall. Wealth inequality rose rapidly in China, Egypt and Hong Kong both before and after the financial crisis. Argentina, India, Korea, Taiwan, Turkey and Russia also experienced a rapid rise over the whole period, although Korea and Turkey had only a moderate rise before 2007, Taiwan showed no trend in the early years, and wealth inequality actually fell earlier in Argentina and Russia. Brazil, the Czech Republic, Indonesia, Israel and the United Kingdom also had significant increases in wealth inequality this century, due almost entirely to rises after 2007. Thus countries with rising inequality are spread quite widely across all regions apart from North America.
At the other extreme, inequality fell rapidly in Poland and Saudi Arabia over the period 2000-14, and a significant reduction was experienced in eight other countries, again widely spread across regions and stage of development. They include Malaysia, New Zealand, the Philippines and Singapore in Asia-Pacific; France in Europe; and Canada, Colombia and Mexico in the Americas.
While there is no clear pattern relating wealth inequality trends to region or to the stage of development, there is something distinct about the G7 countries. Only one of them, the UK, recorded rising inequality over the entire period 2000-14, and only three show an increase after 2007 – France, Italy and the UK. This is unexpected, and interesting, for two reasons. First, income inequality has been rising in these countries and there is heightened concern about wealth inequality as well; yet in most of them, equalisation from 2000 to 2007 was sufficient to offset any subsequent rise in inequality. Second, it appears that wealth inequality did not increase in some of the major countries closest to the centre of the global financial crisis. This result may be explained in part by the fact that the crisis saw the wealthy lose proportionally more than those at lower levels of the pyramid. In some countries that equalising effect still dominates, while in others it has been reversed, partly due to strong market performance since 2009.
Determinants of wealth inequality
Many factors are thought to affect wealth inequality, but their precise impact and relative importance are not well understood. Over longer periods, the level and distribution of wealth in a country will depend on the growth rate of the economy, demographic trends, savings behavior, inheritance arrangements, general macroeconomic trends (such as globalization) and government policies affecting, for example, taxation and pension provision. In the short run, the stock of household assets is relatively fixed, so changes in the distribution of household wealth tend to be driven by changes in asset prices, which can affect wealth inequality because the composition of household portfolios varies by wealth level. Cross-country comparisons are also sensitive to exchange rate movements in the short run.
Wealth inequality and asset prices
The composition of household portfolios tends to vary by wealth level in a systematic way. For middle wealth groups, equity in the family home is often the dominant feature. At lower wealth levels, savings accounts are more prominent – and debts are also more evident – while equity in private businesses and listed companies is heavily concentrated higher up the distribution. As a consequence, stock market appreciation tends to favor wealthier individuals and to cause the share of top wealth groups to increase. This leads to the expectation that wealth inequality will fall when financial assets are declining as a fraction of the household portfolio – as happened during the early years of this century; and to rise when the share of financial assets is increasing – as happened after the financial crisis.
The impact of house price rises is more difficult to assess. The middle class are expected to benefit disproportionately, since owner-occupied housing is more prominent in their portfolio; this has an ambiguous effect on overall inequality, since the top and bottom wealth shares both tend to fall. Furthermore, second homes and investment property form a significant part of the portfolios of wealthier individuals, and appreciate in line with owner-occupied housing. So the impact of house price rises on wealth inequality is not easy to identify and measure. One study for the United Kingdom concluded that rising house prices tend to reduce the share of the top wealth percentile. Another study based on data for the United States echoes this result by showing that the ratio of equity prices to house prices has a powerful positive effect on the wealth share of the top percentile.
Some commentators have claimed that rising equity prices are a consequence – as well as a cause – of rising inequality. It is suggested that rising income inequality in the United States from the 1970s onwards raised the disposable income of the top groups, who typically save a higher proportion of their income. As captured in Figure 3, this led to an increase in funds seeking investment opportunities, driving down interest rates and raising stock prices, which in turn created further capital gains for the top income groups, propelling income inequality to even higher levels. In addition, the fall in interest rates encouraged the housing bubble that developed in the United States in the early 2000s and fuelled the unsustainable growth of debt, which triggered the financial crisis of 2007-08. If this account is even partially true, it raises concerns about the implications of the widespread rise in wealth inequality since 2008, and about the implications for equity markets once low interest rates are no longer regarded as a priority by central banks.
Wealth inequality and wealth growth
Trends in wealth inequality across countries are also likely to depend on the rate at which household wealth is growing. Fast growing economies are usually associated with successful entrepreneurs and the rapid emergence of young businesses owned by a family or a small group of shareholders. The new wealth created when these businesses are established – and later valued and listed – is highly concentrated, leading to an increase in overall wealth inequality. In the longer run, ownership of the new assets is dissipated via broader shareholdings and intergenerational transfers. However, the initial increase in inequality can persist for many years. Effects like these can be seen today in the more rapidly growing transition countries and emerging markets, and help account for the high levels of wealth inequality in emerging markets evident in Table 1.
Social and demographic impact on wealth inequality
In the longer run, wealth inequality will be affected by demographic trends, of which the most important are rising longevity and ageing societies. As the length of retirement is extended, savings for lifecycle purposes will become a more important component of personal wealth. Lifecycle saving is far from equal, and rising income inequality will magnify the differences; but on balance, greater lifecycle saving and pension wealth is expected to reduce wealth inequality. Furthermore, the reduction may well be sufficient to offset any increase in inequality associated with ageing populations, due to the fact that older individuals tend to be wealthier than average.
Other social and demographic factors will also have an impact in the longer term. The spread of popular assets – especially owner-occupied houses, cars and other consumer durables – is seen as a major reason for the secular decline in wealth inequality during the 20th century. Starting from a world in which the working class formed a large fraction of the population and lived in rented accommodation, used public transport and had few possessions, it was inevitable that wealth inequality would decline. Smaller household sizes and more equal treatment of females are also likely to affect trends in wealth inequality, although the precise impact is difficult to assess.
Inheritance and wealth inequality
Property rights and inheritance customs are core subjects in understanding the level of wealth inequality and its transmission over time. In traditional rural societies, wealth inequality is almost synonymous with unequal landholding perpetuated through inheritance. This is true in the Indian subcontinent, parts of Africa, and perhaps most famously in the large latifundia of Latin America, where conditions have historically been semi-feudal. In these circumstances birth dictates opportunity to a large extent. In contrast, the settler societies of Argentina, Australia, Canada and the United States handed out land freely, creating a remarkably equal initial distribution of wealth. Similar impacts occur when land is redistributed on an egalitarian basis through land reforms like those in South Korea and Japan after World War II. In other countries, complex patterns of property rights may mean that the nominal owner of a piece of land may not have the right to work it, or to sell it. Yet other examples exist of countries where land has been redistributed and the more equal distribution frozen through a prohibition on sale – China is the leading current example, but Ethiopia has a similar setup. These complexities pose great challenges to understanding and analysing wealth inequality.
In advanced industrial societies unequal landownership is not a core social issue, but inheritance remains an important route to wealth ownership for some people. Furthermore, inherited wealth tends to be quite unequal since middle and lower income families cannot afford to bequeath much, so children of the wealthy benefit disproportionately. In certain situations, the impact may be equalising. For example, it may assist the spread of new wealth amassed in the first generation fortunes of successful entrepreneurs. Overall, however, it is likely that inheritance tends to raise the level of wealth inequality and to ensure that wealth inequality persists over time, especially in slower growth economies.
Taxation and government policy
Governments can have large impacts on wealth inequality, in a range of ways, some of which tend to be overlooked. High inflation restricts people’s ability to build wealth through saving, and sudden unexpected bouts of inflation can erode or even wipe out the savings of broad groups. A lack of secure property rights can have a similar chilling effect on entrepreneurship or accumulation of real assets. As well as reducing growth rates, such factors can help to generate high wealth inequality. But higher wealth concentration can also result from more benign influences. For example, strong social security programs — good public pensions, free higher education or generous student loans, unemployment and health insurance — can greatly reduce the need for personal financial assets. Public housing programs can do the same for real assets. This is one explanation for the high level of wealth inequality we identify in Denmark, Norway and Sweden: the top groups continue to accumulate for business and investment purposes, while the middle and lower classes have no pressing need for personal saving.
Governments can also reduce wealth inequality, of course. The sheer size of the public sector has an impact. More economic activity undertaken by the public sector leaves fewer opportunities for private entrepreneurship and investment. Progressive income or estate taxes, and taxes on wealth or capital income, reduce rates of return and hamper asset growth. High levels of taxation on large estates appear to be one of the reasons why wealth inequality declined during the 20th century, as wealthier individuals transferred ownership of core assets during their lifetime. Nowadays, family trusts and similar arrangements are frequently used to mitigate estate tax liability, so the impact is now much weaker.
Similarly, while progressive income and capital taxes are likely to lower wealth inequality, flatter tax structures will lead to rising inequality, as some commentators have suggested has happened in recent decades. More positively, tax shelters for retirement saving give the middle class more incentive to accumulate assets. This will tend to reduce the top wealth shares over time, although the shares of bottom wealth holders may fall as well.
Summary and conclusions
We began by looking at the United States, which has good data on wealth distribution and a long time series on wealth inequality. The share of both the top decile and the top percentile declined between 1910 and 1970, and both have trended upwards since then – from 28% to 34% for the top percentile, and from 64% to 75% for the top decile. Trends broadly similar to those for the United States have been recorded for Australia and for eight European countries.
We are reporting for the first time estimates of the wealth share of the top decile constructed on a consistent basis for each year since 2000. The procedures we employ use rich list information to adjust for missing wealth holders at the very top of the wealth distribution. Our results show that wealth inequality varies considerably among developed countries: the share of the top decile ranges from less than 50% for Belgium and Japan, to over 70% for Switzerland and the United States. Among emerging market economies, however, unequal wealth is much more evident: out of the 24 countries we consider, 13 are classed as “very high inequality” with top decile shares above 70%.
As regards wealth inequality trends, our results for the whole period 2000-14 show that wealth inequality rose in exactly half of the 46 countries monitored. Splitting the period reveals markedly different experiences before and after the financial crisis: inequality fell in 34 countries in the earlier years, but in only 11 countries after 2007. This pattern is broadly reflected in the regional experiences, although inequality rose in China and India both before and after the financial crisis, and declined slightly in North America in both sub-periods. Examples of rising and falling inequality are found among developed countries and among emerging markets, so wealth inequality trends show no clear link with the stage of development. However, it is interesting to note that only one G7 nation – the United Kingdom – appears in the list of 23 countries recording an increase in inequality this century.
Many factors contribute to the level and trend of wealth inequality, and their interactions are highly complex. In the short term, asset prices have a strong effect, with the relative fortunes of the wealthy rising and falling with the stock market. It is likely that the abrupt switch from decreasing inequality up to 2007 to increasing inequality in the years after 2007 is linked to the change in the relative importance of financial assets in household wealth, which followed the same pattern. In contrast, house price increases tend to favour the middle class, prompting a reduction in top wealth shares. Interestingly, rising inequality in recent years may have contributed to asset price increases by providing the top income groups with more funds to invest, and caused wealth inequality to rise further, by giving those lower down more reason to borrow.
Over longer periods, wealth inequality is influenced by economic growth, demographics, savings behaviour, landholding, inheritance and government policy. Fast economic growth, for example, is expected to lead to the rapid rise of new businesses, raising inequality. This may account in part for the high level of wealth inequality evident in emerging market economies. Patterns of landholding and the transmission of land from generation to generation is an important consideration in developing countries, while inheritance more generally will tend to support higher levels of inequality, especially in slower growth economies.
Governments can influence the level and distribution of wealth in many ways. Higher levels of taxation – on income, capital, property or inheritance – are all expected to reduce inequality in the longer run, although the repercussions on personal incentives are widely debated. Encouraging wealth creation through tax advantages given to retirement savings programs is less controversial and will tend to reduce inequality. Welfare state policies, including public pensions, help to reduce income inequality; somewhat perversely, however, they reduce the need for lower and middle income families to save, lowering their wealth and tending to raise wealth inequality.
Given the complex sources for wealth inequality it is difficult to predict future changes. However, in China, India and some other emerging market countries, slower growth may bring a deceleration in the rise of wealth inequality, and time will allow recent new fortunes to spread among a wider group of owners. Wealth inequality may also fall as the share of financial assets in aggregate total wealth stabilises or declines. In mature economies, policies to address wealth inequality are receiving increased attention and can hopefully be designed to avoid unwanted effects on growth or economic security. Among emerging markets, policy makers would be advised to study countries, such as Singapore, which have tried to ensure that wealth gains are broadly shared, and which have succeeded in keeping wealth inequality in check.