Economic inequality is a broad concept, which involves the differences found in various measures of economic well-being among individuals in a group, among groups in a population, or among countries [6]. Still, when we read about high “inequality”, what the media usually talks about is Wealth Inequality, which refers specifically to the differences found in the distribution of ownership of the assets within a society [7]. As you can imagine, measuring such gaps is not a trivial matter. Simply defining wealth is a tricky task in itself. A commonly used definition, which is in fact the one used by Oxfam in their research, defines wealth as the difference between the total amount of assets an individual has and his/her liabilities [7] [8]. This definition is one of the first complaints researchers have on Oxfam’s methodology [4]. Assets minus liabilities means that young people are disproportionately poor, just consider the case of a recently graduated student, who does not have any assets (lands, houses or property) but who most certainly has some important debts (college fees). Most of us would agree that a recently graduated young woman has a bright future with important earning potential, but this definition of wealth will likely rank her among the poorest.

There is a reason why Oxfam uses this definition: they have no choice. They have no choice because they do not use their own data for any of the reported results; they rely completely on data published by the Credit Suisse Research Institute. They publish every year the Global Wealth Report [8] a publication where wealth information from many countries around the globe is gathered and analyzed, an exercise that has been replicated since 2000.

According to the Credit Suisse data, if we summed up everything you own, everything your neighbour owns and all assets the whole planet possess, including everything I own, which admittedly is not much, the world had an overall estimated wealth of 280 trillion US dollars in 2017. In order to reach this number, the Global Wealth report estimates the wealth holdings of households from all around the world. To obtain these, researchers first estimate the average level of wealth for each country. This is done on a country by country basis, using whatever information is available to estimate, with emphasis on “whatever”. For 48 countries, the report relies on household balance sheet (HBS) data, although for 25 of these, the information covers only financial assets [8]. For another 4 countries, which include China and India, estimates are based on household survey data, a less than ideal method for estimating income or wealth, as reported by official surveyors. For instance, in Mexico, the National Institute of Statistics and Geography reported consistent problems while using a standardized survey to estimate income in the country [9]. Still, these are the countries where the Global Wealth Report actually has some data. For most of the world, the Credit Suisse Research Institute has no reliable source of information, so estimates of average wealth levels are obtained with modeling techniques, specifically with three regression models aimed to predict non-financial assets, financial assets and liabilities [8]. The last two of the models are estimated as Seemingly Unrelated Regression Models (SUR), which imply that the errors obtained by each of these models are assumed to be correlated. This is not done for the financial assets models, as the authors claim insufficient sample size for it. Finally, for about 44 countries there’s absolutely no information available and their wealth is estimated as the regional average.

After the wealth of each country has been estimated, a second step involves constructing the distribution of wealth holdings within each nation [8]. For 31 countries, survey data is used. For the rest of the world, the authors use the relationship between wealth distribution and income distribution to use models to generate very rough estimates of wealth distribution for countries which have information available on income distribution. For countries where information on the model predictors is not available, the regional average is again used.

These procedures generate a wealth distribution curve for each country. This information is used with a modified version of the Shorrocks-Wan ungrouping program [8], which constructs a simulated sample conforming exactly to any set of values from a Lorenz curve, which is the traditional graphical representation of income or wealth distribution. This produces samples for each country in the world, each with its corresponding sampling weight to match population totals and consistent with that country’s estimated wealth distribution. Samples for all countries of the planet are obtained and then merged in a single, global dataset of about 1.2 million records, which is the one Oxfam works with. Well, almost. A final adjustment is made, based on Forbes’ list of billionaires [8]. They use the number of billionaires reported by Forbes to fit a Pareto distribution to the upper tail of 56 countries and replace the estimated wealth with the new values.

In briefing, to estimate global wealth the Credit Suisse Research Institute uses official government data, household surveys, mean imputation, modeling techniques and even Forbes’ list of billionaires. Obviously, mixing all those sources of information has the effect of drastically reducing the accuracy of the estimation. In fact, just using in the same context survey data from different countries is less than ideal if the survey methodology is not standardized; now consider merging all those different data sources. The authors themselves conclude that the estimation is far from perfect and that revisions are “inevitable” [8].

With this dataset, Oxfam calculates all statistics for their report, at least for the wealth distribution. They use other sources of information for specific parts of the report, such as their estimates of tax evasion, which are based on work by Gabriel Zucman, or the wealth accumulated by billionaires, which is entirely based on Forbes’ list [10].

The Oxfam team is well aware of the liabilities in the Credit Suisse data they use, particularly their wealth estimations for certain countries that tend to be way off. The reason is the heterogeneous data sources the Credit Suisse report uses, which produces not so accurate results for particular regions. This is the reason why Oxfam made such abrupt shifts in their estimations of wealth for the bottom 50% of the world from 2016, which led to the increase in the estimated number of billionaires who accumulated enough wealth to match the fortunes of the bottom half of the planet. Initially in their 2016 report, Oxfam published that it required only the 8 richest persons in the world to match the fortune of half of the planet [4]. But this number assumed that the total accumulated wealth of the bottom 50 percent was of $409 billion. This year’s Credit Suisse report uncovered about $8 trillion of global wealth in 2016 that was previously not counted and more than $1 trillion of it belonged to the poorest half of the planet, meaning that their accumulated wealth actually more than tripled [5]. Most of the “discovered” assets came from India, China and Russia, places where household survey data did not accurately measure wealth. After Oxfam considered this additional money, they revised the number of rich people required to match the wealth of the bottom half of mankind to 61. This actually means that the world was not in such a bad condition as we thought in 2016: average net wealth for the bottom 50 percent is no longer $110 per person, as Oxfam first estimated, but $427 per person [5]. Whether this year’s numbers also failed to consider trillion of dollars of fortunes is another matter entirely and we probably won’t find out until next year’s numbers are published and this year’s revised.

With this in consideration, it is no surprise that Oxfam’s yearly report is viewed with skepticism by a fair share of the scientific community. But considering the immense challenge that represents gathering income and wealth data for all countries on earth, it might not be appropriate to immediately discredit Oxfam’s effort. The Credit Suisse data has many opportunities to improve but it is still the only resource available to offer global estimates of wealth. Besides, Oxfam data is often in line with other research done on wealth inequality, at least within countries. Pew Research Center published in 2017 that wealth gap between upper-income families and lower- and middle-income families reached the highest levels recorded, based on data from the Federal Reserve Board’s Survey of Consumer Finances [11]. As I mentioned, measuring wealth with surveys is incredibly hard, as people tend to under-report their assets or, worse, tend to not participate at all. Usually upper-income households are harder to reach with surveys, which is another issue. But even with all those caveats, survey data in the U.S. agrees with Oxfam’s conclusions. Stanford’s Center on Poverty & Inequality also provides yearly analyses of trends in income and wealth inequality for the U.S., this time based on annual updates to tax data [12], with similar results. Even more, a closely related indicator, income inequality, has shown signs of going upward in many parts of the world, at least according to data from OECD countries [13]. Their estimates reflect that the average income of the richest 10% of the population is about nine times that of the poorest 10% across all OECD countries, the largest gap for the past half century.

And to be fair, it is certainly true that Oxfam’s greatest interest lies not in providing the most accurate estimates of wealth distribution, but to bring the issue of global wealth inequality to light and motivate a broader discussion on the economic issues and policies that could help drive inequality down. Since no one actually knows how to do it, as the causes of wealth inequality are a topic of intense debate even today. Education has been identified as an important factor that drives inequality: people with lower education levels have their opportunities restrained and their wealth is also affected in consequence [14]. Low-skilled functions tend to be associated with lower positions in the wealth distribution as well, as competition for these positions has decreased, thus exacerbating the effect of education on the wealth gap. Environmental factors, such as the political order, tax policy or stagnant wages are also major drivers for inequality. But even though most economists agree on the influence of these factors, there is ongoing debate on the actual mechanisms that explain the relationships between these variables and inequality itself.

As such, Oxfam’s report is far from being a perfect statistical exercise, but it is the best available resource we have for a global comparison. And even as imperfect as it is, it has certainly been successful in its other goal: bringing the inequality issue to the debate table.

REFERENCES:

[1] Reward Work, Not wealth Oxfam International (Jan, 2018)
https://www.oxfam.org/sites/www.oxfam.org/files/file_attachments/bp-reward-work-not-wealth-220118-en.pdf?cid=aff_affwd_donate_id78888&awc=5991_1516715345_0a84322c20ef396277dc8ed070020d3e

[2] Hope, Katie ‘World’s richest 1% get 82% of the wealth’, says Oxfam. BBC News (Jan, 2018)
http://www.bbc.com/news/business-42745853

[4] Hope, Katie Eight billionaires ‘as rich as world’s poorest half’ BBC News (Jan, 2017)
http://www.bbc.com/news/business-38613488

[3] The World’s Richest 1% Took Home 82% of Wealth Last Year, Oxfam Says. Fortune Website (Jan, 2018)
http://fortune.com/2018/01/21/oxfam-report-global-inequality-billionaires/

[8] Global Wealth Report 2017. The Credit Suisse Research Institute Website (November,2017)
https://www.credit-suisse.com/corporate/en/research/research-institute/global-wealth-report.html

[5] Murphy, Tom. Why The Internet Loves And Hates Oxfam’s Global Inequality Report. NPR Website (Jan, 2018)
https://www.npr.org/sections/goatsandsoda/2018/01/24/580087309/why-the-internet-loves-and-hates-oxfam-s-global-inequality-report

[9] Barragán, Daniela & Dulce, Olvera. El 63.3% del ingreso en México se concentra sólo en el 30% de los hogares, dice encuesta del INEGI. (Available in Spanish) SinEmbargo website (Aug, 2017)
http://www.sinembargo.mx/28-08-2017/3294496

[6] Economic Inequality. Wikipedia, The Free Encyclopedia (Last edited March, 2018)
https://en.wikipedia.org/wiki/Economic_inequality

[7] Distribution of Wealth. Wikipedia, The Free Encyclopedia (Last edited March, 2018)
https://en.wikipedia.org/wiki/Distribution_of_wealth

[14] Income Inequality. Investopedia Website (2018)
https://www.investopedia.com/terms/i/income-inequality.asp

[10] Reward Work, Not Wealth. Methodology Note. Oxfam International (Jan, 2018)
https://oxfamilibrary.openrepository.com/oxfam/bitstream/10546/620396/42/tb-reward-work-not-wealth-methodology-note-220118-en.pdf

[11] Kochhar Rakesh et al. How wealth inequality has changed in the U.S. since the Great Recession, by race, ethnicity and income. Pew Research Center (Nov, 2017)
http://www.pewresearch.org/fact-tank/2017/11/01/how-wealth-inequality-has-changed-in-the-u-s-since-the-great-recession-by-race-ethnicity-and-income/

[12] Research Projects. Stanford Center on Poverty & Inequality. (Stanford University, 2017)
https://inequality.stanford.edu/research/projects#income-wealth

[13] Inequality. OECD – Better Policies for Better Lives. OECD Centre for Opportunity and Equality. OECD. 2018
http://www.oecd.org/social/inequality.htm