An important finding: Income inequality is neither ubiquitous nor inevitable.


Rising levels of inequality has become a key political issue in recent years; “the defining issue of our time,” as Barack Obama described it back in 2013. It was at the heart of the Occupy movement protests, and has received a huge amount of attention in the media and in policy circles. But much of this discussion has focussed on what’s been happening in rich countries, in particular upon trends seen in the US. We tend to hear less about inequality in the rest of the world.

How has income inequality within countries been changing around the world more generally?

To answer this question, we brought together estimates of income inequality for two points in time: today and a generation ago in 1990. Our metric of income inequality is the Gini index – explained here – which is higher in a country with higher inequality.1

We rely on estimates from two online databases: PovcalNet, run by the World Bank, and the Chartbook of Economic Inequality, which I published together with Tony Atkinson, Salvatore Morelli, and Max Roser.2 This gave us a sample of 83 countries, covering around 85% of the world’s population (download the data here).

The chart below compares levels of inequality today with those one generation ago. If you’ve not seen this sort of chart before, it may take a moment to understand what’s going on. How high a country is in the chart shows you the level of inequality in 2015. How far to the right shows you the level of inequality in 1990.3 In addition, a 45 degree line is plotted. Countries below this line saw a fall in the Gini index between the two dates; countries above saw an increase.

Four observations on inequality across the globe

So what does the chart tell us about inequality within countries across the world?

  1. No general trend to higher inequality: It’s a mistake to think that inequality is rising everywhere. Over the last 25 years, inequality has gone up in many countries and has fallen in many others. It’s important to know this. It shows that rising inequality is not ubiquitous, nor inevitable in the face of globalisation, and suggests that politics and policy at the level of individual countries can make a difference.
  2. Note the diversity between countries: As well as there being different trends, notice how very different the level of inequality is across countries. The spread you see – from the highest inequality countries in Latin America and Sub-Saharan Africa, to the lowest-inequality countries in Scandinavia – is much larger than the changes in individual countries over this period.
  3. There are clear regional patterns: To bring this out in the chart you can highlight particular regions by clicking on the labels in the legend on the right side.

    – Almost all Latin American and Caribbean countries show very high levels of inequality, but considerable declines from 1990 to 2015.

    – Conversely, advanced industrial economies show lower levels of inequality, but rises in most, though not all, instances.

    – A number of Eastern European countries experienced rising inequality as they transitioned from socialist regimes.4

    – Across the six countries in our sample from the Middle East and North Africa region, we mostly see falls. In Sub-Saharan Africa and East Asia and Pacific, the trends are more mixed.

  4. Across countries, the average level of inequality has not changed:

    – The rises and falls seen in the Gini index in different countries more or less cancel out, the average Gini across countries fell marginally from 39.6 to 38.6.

    – There were rises in inequality in some of the world most populous countries, including China, India, the US and Indonesia (together accounting for around 45% of world population). As a result, if we weight countries according to the size of their population we see that this weighted average Gini index increased by four percentage points, from 36.7 to 40.8.

    – This means that, whilst in terms of the average country the Gini index stayed roughly constant across the two periods, the average person lived in a country that saw rising inequality.

  5. Levels of inequality are converging: Interestingly, the chart shows that there was some convergence in inequality levels across countries over the last 25 years. Amongst those countries with a Gini index below 40 in 1990 (left half of the chart), hardly any saw substantial falls to 2015. Amongst those with a Gini index above 40 in 1990 (right half of the chart), hardly any saw substantial rises. As already pointed out, this apparent convergence works largely through regional dynamics. That said, those countries in our sample from the Sub-Saharan Africa and East Asia and Pacific regions are more evenly split between rising and falling levels of inequality, but still roughly fit this convergent pattern.

 

Conclusion

The Gini index is just one of the many ways we can measure inequality, each with their own pros and cons. You can read more about this in the ‘Further notes’ section below, as well as in our entry on Income Inequality.

Nevertheless, it is clear from the chart we cannot make generalisations about inequality across the globe based on what we see in rich countries. Nor should we limit ourselves to thinking inequality must either be going up or going down, full stop. Posing the question in such a polarised way precludes a meaningful answer. Whether inequality is rising or falling depends on where, when and what aspect of inequality we have in mind.

But this is very important to know in itself. It shows us that rising inequality is not just an inevitable outcome of global economic forces, completely beyond our control. National institutions, politics and policy play a key role in shaping how these forces impact incomes across the distribution. Being attentive to the differences between countries is an important step in knowing what can be done to reduce inequality.

Further notes on data quality and interpretation

In the chart above, we’ve used used the Gini index, as estimated from household survey data, as our measure of inequality. In this section we address a number of downsides to this approach.

Our main reason for using the Gini index was the wide range of countries for which it is available. However, in order to have this coverage, we had to put together estimates relating to a very heterogenous mix of survey methodologies and concepts. In particular, surveys of both household income and consumption are included. Incomes are, in general, distributed more unequally than is consumption expenditure and thus the Gini estimates are not entirely comparable. For instance, Latin American and Caribbean countries (which generally use income surveys) may appear more unequal in relation to Sub-Saharan African countries (which generally use consumption surveys) than they really are.6

There is also a variety of income concepts in operation, as well as other inconsistencies. The estimates for advanced economies taken from the Chartbook of Economic Inequality mostly relate to disposable household income, equivalised to account for household composition. The income concept in the estimates from Povcal is unknown, but most likely varies considerably across countries. Moreover, the figures in Povcal relate to per capita household income, without adjusting for household composition. This most likely overstates inequality relative to estimates based on equivalised income.

To mitigate these comparability issues somewhat – at least in terms of trends over time – only countries where estimates based on the same broad welfare basis (income or consumption) around both reference years were included. The source of the estimates was also kept constant over both years.

However, irrespective of these comparability issues there are a number of more general concerns about measuring inequality using the Gini index based upon household survey data. One major issue is that the incomes of the very richest are often missing or under reported in household surveys, thereby understating the degree of inequality. The online database WID.world provides data based on an alternative approach that considers the share of pre-tax income received by the top 1 or top 10%. The approach more reliably captures top incomes by incorporating tax data and national accounts aggregates.

For some countries, the trends over the last 25 years look less benign when we consider top income shares, which focus on movements at the top of the distribution, without taking into account the redistribution effected by governments through taxes and transfers. The most notable instance of this is Russia. The Gini index, based on consumption expenditure, shows a fall of around 10 percentage points between 1990 and 2015. At the same time, the share of income received by the top 1 % more than doubled, from 7 to 20 percent.

Only 22 of the 83 countries in the chart also have top income share estimates around both reference years, preventing a full comparison. But of the three countries that showed a fall in the Gini index of more than two percentage points, and were also available in the WID.world data, two (Russia and Malaysia) showed a rising top 1% income share over the same period.7

More generally though, it is important to realise that there are many ways of measuring inequality and that trends can look very different depending on which approach one takes. Hoy (2015), using very similar data and approach to that used here, shows that absolute inequality within countries went up on average in every world region over a similar period.8 But it would be unhelpful to think in terms of one approach giving the ‘true’ answer and another being misleading. From what empirical evidence is available, both relative and absolute inequality matter to people.9 Having multiple perspectives on inequality is crucial if we are to capture people’s varied intuitions about what inequality consists of.

Footnotes

  1. Whilst we refer to ‘income inequality’ here, the estimates in the chart are based upon household surveys of both income and consumption, depending on the country. This is discussed more in the ‘Further notes’ section below.
  2. For China, we take the series presented in table 1.B from Kanbur, R, Y Wang and X Zhang (2017) “The great Chinese inequality turnaround“, CEPR Discussion Paper No. 11892.
  3. The estimates are based on household surveys of either income or consumption. This heterogeneity, as well as that between varying definitions of income and survey methodologies used introduces important inconsistencies that may affect the levels or trends seen across countries. See “Further notes” below.
  4. Given that estimates are not available for every year, we selected those falling closest to 1990 and 2015, up to a maximum of 5 years’ distance.
  5. The full time series of estimates for these countries shows that the rises were mostly concentrated in immediate post-Soviet period in the 1990s.
  6. Amongst OECD countries, Noland et al. (2017) note, in general, larger rises in inequality since 1980 in those countries beginning from a lower starting point. “Inequality and Prosperity in the Industrialised World“. CitiGPS: Global Perspectives & Solutions. September 2017.
  7. Alvaredo and Gasparini (2013) settle on a downward adjustment of around 15 percent to Latin American and Caribbean estimates in their analysis of the Povcal data. This means a Gini index of 50 is reduced to around 43. They base the figure on the average difference between income and consumption Gini estimates for those years and countries (within Latin America and the Caribbean) in which both were available.
  8. Of the 14 countries in the WID data which saw a rise in the Gini index of more than 2 percentage points, there countries saw a fall in the top 1% income share of which only one – Cote d’Ivoire – saw a fall greater than half a percentage point.
  9. Hoy, C. “Leaving No One Behind: the Impact of Pro-Poor Growth“. ODI Report. October 2015.
  10. Absolute inequality measures capture increases in absolute, rather than relative differences between people’s incomes. If the average income of the top 10% is $100,000 and the average income of the bottom 10% is $10,000 then the absolute difference between the groups is $90,000. Doubling everyone’s income would not increase relative inequality, but it would increase absolute inequality.
  11. There have been a small number of surveys involving university students about this question. In each case, about half of the respondents showed they thought about inequality in absolute terms. Martin Ravallion summarises these studies in Ravallion, M. (2014). Income Inequality in the Developing World. Science (New York, N.Y.). 344. 851-5. 10.1126/science.1251875


This post was previously published on www.ourworldindata.org and is republished here under a Creative Commons license CC BY-ND 4.0.

What’s your take? Comment below or write a response and submit to us your own point of view or reaction here at the red box, below, which links to our submissions portal.

submit to Good Men Project

◊♦◊

Sign up for our Writing Prompts email to receive writing inspiration in your inbox twice per week.


If you believe in the work we are doing here at The Good Men Project, please join us as a Premium Member, today.

All Premium Members get to view The Good Men Project with NO ADS.

A $50 annual membership gives you an all-access pass. You can be a part of every call, group, class, and community.
A $25 annual membership gives you access to one class, one Social Interest group, and our online communities.
A $12 annual membership gives you access to our Friday calls with the publisher, our online community.

Register New Account

Log in if you wish to renew an existing subscription.

Choose your subscription level

Credit / Debit Card
PayPal

By completing this registration form, you are also agreeing to our Terms of Service which can be found here.

 

Need more info? A complete list of benefits is here.


◊♦◊

Get the best stories from The Good Men Project delivered straight to your inbox, here.

 

Photo credit: Istockphoto.com

The post An important finding: Income inequality is neither ubiquitous nor inevitable. appeared first on The Good Men Project.

Leave your comment