Photo: Pau Casals / Unsplash
Photo: Pau Casals / Unsplash

Global poverty: fantasy or reality?

Unpacking the recent Smith and Hickel debate on the evolution of global poverty.

Rishabh Kumar teaches economics at the University of Massachusetts Boston. He can be reached on Twitter at @Kumar_EconIneq.

Global poverty has been declining for a few decades, or has it? A few months ago, this question was at the heart of a debate stirred by Noah Smith, an economics commentator for Bloomberg. Channeling those who argue poverty is less today than at any other point in history, Smith wrote a newsletter on the online platform Substack criticising the loudest voice on the other side – Jason Hickel. For some time, Hickel (an economic anthropologist based in London) has been arguing that global poverty reduction is a fantasy promoted by institutions like the World Bank. This story is then marketed by wealthy Davosians like Bill Gates and intellectuals like Steve Pinker to make everyone feel better about the world as it is and as it has been progressing for the last few decades. I am tempted to believe that most economists agree with the viewpoint adopted by Noah Smith. When Smith published his article, it generated plenty of support on Twitter – the best proxy for a crowd in the age of pandemia.

The problem facing the poor today is that the countries they reside in are no longer following a typical development path.

There is a simple explanation for why Hickel stands alone on this stage – it is a difficult task to show evidence that global poverty has not declined. Log into the World Bank's database, and you will be hit by a promotional graphic on the homepage, which shows that a greater fraction of the world now lives above the poverty line than at any other point in history. The poverty line is a global monetary threshold, measured in International Dollars (i.e., adjusted for variation in cost-of-living). It represents a pooled aggregate from national estimates of the cost of purchasing the basic needs of human life (food, housing, energy, education, health, etc.). In the 2015 estimate, less than 10 percent of the world lived below this threshold compared to almost 43 percent in 1981. After the economic consequences of the pandemic, this number has likely increased but is unlikely to return to past levels and is related purely to a shock (rather than systematic swelling).
The World Bank also adds a notion of transparency by providing open access to simple tools, allowing users to manipulate the poverty line; the results do not change. This figure is a testament to the fact that the most brutal elements of history are being left behind by today's world – falling infant mortality, rising life expectancy, and ultimately, people consuming at a higher standard in modern society.
Last year, the magazine Foreign Affairs published a debate between foreign aid stalwart Jeffrey Sachs and the super-duo of development economics, Abhijit Banerjee and Esther Duflo, on the best means to achieve zero poverty. Still, no one questioned the statistical summary supplied by the World Bank. Data is empty without a story, of course. The interpretation is that poor countries were catching up with advanced nations by growing faster. In the process, millions were lifted out of poverty because at least some of those gains surely went to the poorest strata of developing countries.

There is a glaring disconnect between the macro-theory of poverty reduction and the dedicated field of poverty measurement.

From a Rawlsian lens, the world is better today because the poorest are better off. To add a dose of moderation amongst the optimists, it is also standard to patronise indigenous industrial policy in developing nations as resilient to the excess of a globalised system of production and exchange. To take this route, the goalposts shift subtly from poverty to global inequality (the gap between rich and poor countries) making trickle-down an international phenomenon. For example, Noah Smith critiques Hickel's pessimistic outlook on capitalist growth and globalisation by sprinkling a few ambiguous growth stories which led to lower poverty. India, Smith says, implemented a developmental state in the 1980s (the cited literature says politicians gave up on raising economic growth, bequeathing responsibility to domestic capitalists) and became a prosperity beacon for the world. Currently, less than 10 percent of the country's labour force has formal sector jobs, but the richest Indians are now closer to their global counterparts, even residing in a stratospheric neighbourhood called "Billionaires row."
To better understand the cracks in the optimistic (low poverty) argument, let me frame this question differently: do global poverty estimates reflect the facts about global economic growth? It always strikes me that there is a glaring disconnect between the macro-theory of poverty reduction and the dedicated field of poverty measurement. Lifting people's living standards requires poor countries to follow a very specific path of catch-up growth. This entails drawing on their large surplus rural labour force towards a modern industrial sector where capital enriches their productivity. Labour and capital thus bootstrap their way to higher average incomes. The canonical example is the transfer of informal agricultural workers to organised manufacturing. The latter has low skill requirements, thus capitalising on a resource that developing countries have in abundance. Secondly, manufacturing output can be sold in world markets and is not constrained by demand levels at home. The East Asian miracle economies (Taiwan, South Korea) followed this type of path out of poverty and into high-income status.

Successful stories of development have been peppered with path dependence – geography, natural resources, etc.

The problem facing the poor today is that the countries they reside in are no longer following this development path. Instead, as Harvard University's Dani Rodrik argues, most countries are entering a stage of premature deindustrialisation. According to this framework, poor countries' employment generation in manufacturing has stalled. Economic growth is instead being driven by services. In Southasia, the service sector accounts for almost 50 percent of gross domestic product (averaged across constituent countries) – a level close to the share in richer countries with the caveat that it has happened at much lower per-capita income levels. To be sure, there is no special law that endows manufacturing-led growth as the sole path out of poverty. Nevertheless, Rodrik also argues that, worryingly, the transition to services occurs in the informal sector – from agriculture to small, unorganised service provision in cities. Workers move from the countryside to cities, but their skills are inadequate for employment in formal service sector jobs such as Information Technology or teaching middle school students. Thus, the growth of the service sector represents, to a great extent, the influx of migrants who find themselves working as roadside barbers, street snack sellers, security guards, domestic help, etc., in cities where the cost of living is higher than the village. Work is precarious, and incomes are often earned on a piecemeal basis – when offices are closed on the weekend, the local street food vendor has no goods to sell and accrues no income. One can imagine why even low-wage annual contracts at a manufacturing location may provide a more guaranteed income.
How is this related to poverty reduction? Undoubtedly, the countryside probably provides little opportunity besides disguised unemployment on the farm. But cities do not provide much more either because these workers stay locked into precarious work without benefits, vacation and healthcare. Low-income countries have little state capacity and, thus, provide inadequate social security. Still, the data suggests that at least fewer now live on USD 1.90 or less per day. This part might well be true. But the statistic is incomplete because it is unclear whether this number represents living standards every day for an entire year. Poverty estimates show declining headcounts over time, giving the impression that few live on USD 1.90 per day over all of 2015 versus (say) 1981.

In Southasia, the service sector accounts for almost 50 percent of gross domestic product – a level close to the share in richer countries with the caveat that it has happened at much lower per-capita income levels.

This point is important because the structure of work and income security for household plays a crucial role in determining the extent and duration of poverty. The World Bank aggregates data from national surveys conducted by member countries. The frequency of surveys varies (annual, quinquennial), and implementation is arduous and expensive. In these data collection exercises, surveyors make a daily estimate of consumption based on whatever the household can recollect over a brief period (a week or month at most). To translate these into annual figures, these numbers are then multiplied as if the survey period represents the entire year. In essence, a household's expenditures are a function of its memory. They recall recent expenditures well, and some margin of imperfection permeates longer recall periods. Depending on the type of household surveyed, estimates imputed from recall make a difference. A college-educated employee knows their annual income and spending patterns. Seasonal farmworkers also know the agricultural cycle.
But for households whose income depends on informal contracts, there is likely to be substantial uncertainty. Domestic helpers do not get a benefits package, and nor do they have an employer-sponsored vacation. If we assume that they return home for vacation, essentially, this is unpaid leave; even sick days are mostly unpaid days. Many workers enter into a piecemeal agreement with the employer to be paid for the actual instance of work (service provision) versus a fixed contract covering working days over a pre-specified period. For example, if a self-employed barber has to rush to their village to attend to a relative, they are essentially unemployed for the duration of that visit. The point is that precarious work provides fluctuating living standards, and consequently, spending on the essentials of life is unknowable with full precision.

Data is empty without a story.

The gist of this argument is that the facts of growth put a probability rather than a 100 percent certainty on the annualised imputation of the USD 1.90 per day statistic. Let me show this in a simple numeric example. For a person to be above the poverty line in a given year, their annual income (which approximates hand-to-mouth consumption) is USD 1.90 times 365, which comes to USD 693. However, if we assume that they only find work for 300 days, then the corresponding figure becomes USD 570. Further, if this person is sick for 15 days of the year, the days of work fall further to 285. Converted back to a daily wage, this comes to less than USD 1.50. This estimate is only a rough calculation, but it underlines my point – incomes and living standards at the lower end of the income distribution of poor countries are imprecisely measured and thus, at the margins, poverty headcounts are quite possibly inflated. This manner of growth still provides an escape from poverty, and certainly, democratic states have done much to reduce infant mortality and education, even in rural areas. But it also implies that a large population depends on uncertain earnings in a more expensive location (the city). In the village, the pressure on housing may be less binding because of multi-generational living arrangements even if existence is close to subsistence. The crux of this argument still holds.
It makes sense to ask why these countries did not follow the path shown by East Asia, and more recently, China. Going back to Smith's criticism of Hickel, the former admonishes the latter for excluding China in his set of countries where poverty reduction is not happening. After all, why exclude 20 percent of the world's population from a successful model of exiting poverty? This is a fair point, but it also answers the question I just mentioned. Following East Asia, China, with its vast population, took a head start in becoming the world's manufacturer in the late-20th century.

Factoring in precarity puts into question the idea of comparing the incomes of formal employees with those in informal jobs in one unified, clean dataset which is the 'killer app' of poverty reduction.

To a great extent, this also suggests the possibility of other poor countries getting crowded out from this path by an established leader. Successful stories of development have been peppered with path dependence – geography, natural resources, etc. The argument is not to exclude or include China, but rather China's methodology of escape from poverty itself. This argument becomes especially binding because rich countries, themselves experiencing a structural slowdown, can only provide finite demand for manufacturing demand from labour surplus countries. If indeed the path out of poverty is unique and the Chinese constraint is relaxed, then Bangladesh, currently manufacturing its way to greater prosperity than other Southasian countries, might well serve as a future case of optimism. But given its population size (2 percent of the world), this process will take time to show up in poverty figures at the Southasian and worldwide level.
Meanwhile, in India, one only has to think about the most splendid tech campuses employing workers in savvy exportable services – IT and business services – who offer services to the world. Those working in these offices are not the ones that the World Bank claims are escaping poverty. The latter group is the people working outside these campuses, in the periphery and the homes of these white-collar employees, providing food and other services to these employees. The certainty of uncertain work was made clear by the pandemic-induced lockdowns when the world's media splattered images of exhausted and hungry migrants waiting for transport to take them back to the countryside. At that moment, no one would believe that these people had the security of formal contracts which would provide incomes to survive the lockdown. Ultimately, factoring in precarity puts into question the idea of comparing the incomes of formal employees with those in informal jobs in one unified, clean dataset. That dataset is the 'killer app' of poverty reduction.
All this is to say that the 'madman' of global poverty, as Hickel is sometimes caricatured to be, may have had a reasonable point. As economists, we often forget just how fragile our data are. If the strength of a critique is just the obviousness of publicly available data, then critics should know what that data counts, and does not count. Thinking this through, maybe it is just not that obvious.
Loading content, please wait...
Himal Southasian