the failure to count is directly related to the power of those involved. At the bottom, those who are excluded tend to be from already marginalized groups. The established weaknesses of these data include the almost universal absence of good statistics on lesbian, gay, bisexual and transgender (LGBT) populations and persons living with disabilities, as well as country-specific failures around indigenous populations and racial and ethnolinguistic groups. And the absence of good statistics can lead, in turn, to the absence of a profile for these groups in public policy discussions and prioritization decisions.
Sometimes, supporters of the status quo will use the fact of a phenomenon being uncounted as an argument against addressing the underlying inequalities. The fight for international tax justice, for example, has seen a number of phases. First, policymakers would deny that tax avoidance and tax evasion were a significant problem. Then, when presented with estimates of substantial scale, the response would often be that the big numbers are not sufficiently robust. Only under the pressure that stems from media coverage and public awareness of those big numbers have international institutions themselves begun to estimate the scale of the problem – often coming up with bigger numbers than had campaigners, and finally leading to sustained policy engagement.
There is important value to the process of testing and probing estimates, both to improve their quality and as part of the social legitimation of their construction. But this can also often provide cover for defenders of the status quo simply to raise doubts about the validity of the underlying concerns. In the tax justice space, this is typically expressed as the view that the big numbers are not robust, so there is no ‘pot of gold’ to be had from stopping multinationals’ tax avoidance. It is clearer now, when that view continues to be expressed despite a range of new data sources and research by independent academics and from international institutions, that much of the support for it is politically rather than technically motivated. But in the early stages of the tax justice movement, there was a genuine risk that the absence of better data could have provided the defenders of the status quo with a conclusive argument against progress.
Bad statistics used to advocate for change can be exploited by what we might call the ‘uncounted lobby’. A paradigmatic example has been the claim made that the informal settlements in Kibera, Nairobi, constitute ‘Africa’s largest slum’. Over time, the challenges to this deeply flawed statistical claim have contributed both to improve the true understanding of the area and its development challenges, and to raise the standard to which development nongovernmental organizations (NGOs) hold themselves in respect of similar claims. But at the same time, the absence of good data backing this particular claim has provided fodder to those with a broader political agenda to downplay concerns about global poverty, and to attack NGOs and others arguing for justice.11 The claim was comprehensively debunked by Kenya’s census, and covered with high profile in the Daily Nation newspaper as the results were released in 2010. But two years later, international reporting was claiming credit for applying their scepticism to ‘Africa’s propaganda trail’ – a much wider claim that seemed intent on undermining a whole sector.
Or consider the ‘zombie stats’ around women’s inequality: in particular, the claims that women make up 70 per cent of the world’s (extreme income) poor, and that women own only 1 per cent of the world’s land.12 For the uncounted lobby, the fact that neither of these can be stood up by available data is evidence that supporters of women’s equality are misguided, extreme, talking about a problem that doesn’t exist, etc. For the rest of us, the fact that we (still) don’t have the data to know how extreme the income and wealth inequalities facing women are, is itself an obvious part of the problem.
Being counted does not guarantee that inequalities will be addressed. But being uncounted certainly makes inequalities less visible, and progress less likely. James Baldwin put it better: ‘Not everything that is faced can be changed, but nothing can be changed until it is faced.’13
At the top, inequality is hidden in three main ways. First, inequality is hidden through missing data: while the poorest groups are underrepresented in surveys, high-income households are much less likely to respond to surveys and are therefore omitted. This can be fixed by using data from tax authorities, where available, which has been seen to add significantly to observed inequality.
Second, and most blatantly, there is deliberate hiding of income and assets by those with most power. This means that the tax data itself contains important omissions. Through the use of anonymous companies and the exploitation of bank secrecy, substantial sums of wealth and the resulting income streams are hidden overseas by high-income households all around the world. Thomas Piketty, in his landmark study Capital in the Twenty-First Century, suggested that an estimate of the volume of undeclared assets of around 10 per cent of global gross domestic product (GDP) should be thought of as the lowest possible level. This is part of Piketty’s reasoning for a global wealth tax – that, actual revenue and redistributive effects aside, such a policy would ensure that the distribution data at least exist.
Multinational companies use similar secrecy mechanisms, coupled with accounting opacity, to shift massive volumes of profits out of the tax jurisdictions where they arise. Together, individual and corporate tax abuses drain hundreds of billions of dollars in revenue from governments around the world each year, undermining the effectiveness of progressive, direct taxation – and broader attempts to hold accountable these largest of the world’s economic actors.
The third way in which income inequality goes uncounted is more subtle, but perhaps equally poisonous. The Gini coefficient, the default measure for inequality, is inherently flawed – and in such a way that it is relatively insensitive to the tails of the distribution (the parts we care about most), and increasingly insensitive at higher levels of inequality (the times when we care most). So what is presented as a neutral, technical measure in fact imparts a serious bias to our understanding of inequality and its development over time. Overall, we know much less about inequality of wealth and income at the top than we do for most of the distribution – and the most common inequality measure exacerbates this problem still further. Metrics, as well as statistics, are distorted by power and in turn distort political outcomes.
Similar problems play out in the area of what we might call ‘privately counted’ data. These include cases where data of significant public value is indeed gathered but is held by private companies with an interest in what part of it is publicly countable, and how. The roles of Facebook and Cambridge Analytica in the UK’s Brexit referendum and the election of Donald Trump as US President highlight the potential threat to democracy.14 Another example is that of the limitations of access to medical trial data generated by pharmaceutical companies. Ben Goldacre has shown how the failure to regulate for effective access to medical trial data results in inferior medical treatments continuing to be widely marketed and used – with human impacts up to and including large-scale excess mortality.15
Cathy O’Neil’s analysis of the biases in algorithms reveals just how far we are from the dream that big data or its use in the public or private sectors could be a tool for equalizing power. Instead, there are multiple, opaque avenues for discrimination, both deliberate and accidental, with grave implications for a range of inequalities.16 More trivially, unless you take seriously Bill Shankly’s suggestion that football surpasses life and death in importance, the Tax Justice Network’s Offshore Game project has shown how the unregulated financial secrecy used by owners can lead football fans to suffer all sorts of risks – including exploitation and even liquidation of their clubs.17
The common thread across all the cases touched on here is the relationship between power, inequality and being uncounted – a relationship that demands we pay much more attention to who and what are and are not counted. The sociologist William Bruce Cameron observed that not everything that can be counted counts, and not everything that counts can be counted.18 While this antimetabole is undoubtedly true, so, too, is another: much that goes uncounted matters, and much that matters goes uncounted.
There
