The opposite of enough
30 years of gains on global poverty have been erased by the pandemic. What can we do about it?
… almost every time we have opened our mouths or put pen to paper, in fact, we have celebrated decades of historic progress in fighting poverty and disease.
But we have to confront the current reality with candor: This progress has now stopped.
Bill & Melinda Gates, “The Gates Foundation - COVID-19: A Global Perspective”
Yesterday, Argentina released its poverty statistics for the first six months of 2021. They were dismal, and showed a similar level to the one from a year ago, which I’ve talked about here in the long distant past. This isn’t an outlier: global poverty has gotten worse everywhere on the planet for the first time in nearly 30 years. What can we do about it?
30 years of good news, one year of bad ones
As controversially tweeted by Bill Gates two years ago, global poverty has gone down significantly in the past 200 years - particularly in the last 30 or 40. In 1990, about 2 billion people lived in extreme poverty (under 1.90 dollars a day); by 2015, just 750 million did so. If you look at other, higher estimates of a global poverty line (say, 3.2 dollars, 5.5 dollars, or 7.4 dollars) then the story is similar, if a bit less impressive.
This doesn’t tell all the story, but it tells enough of it to give us a clear image. Just like GDP per capita, (a measurement of) extreme poverty is highly correlated with several indicators of welfare that have been improving signifcantly as well: hunger, life expectancy, child mortality, infectious diseases, violent crime, literacy, etc. This goes in line with research showing that, in terms of growth, poor countries are finally catching up (as I talked about long ago, too). This process started at the end of the 60s, during decolonization, and went into hyperdrive during the early to mid 90s - precisely when economists started discussing the concept, entertainingly enough. The champions of this new era of convergence where definitely South and Southeast Asia: back in 1990, over 58% of people lived in extreme poverty, the same percentage as in Subsaharan Africa. Today, just 1.5% do - better than Latin America. The worst kind of poverty imaginable declined by double digits almost every year of the last decade.
All of this seemed good: poor countries were getting richer (with some, like Malaysia, on their way to actually become developed), global poverty down, development indicators up. The Sustainable Development Goals, a really ambitious set of wellness targets, seemed somewhat possible - remember, more than two billion people lived in extreme poverty (i.e. under the 1.9 dollar line) back in the 90s.
And then COVID happened. Before 2020, we expected 31 million people to fall out of poverty last year, a modestly successful figure in line with previous years. By June of this year, we realized that in fact 97 million fell into it - the worst figure since the late 90s, and 127 million more than predicted. Put another way, we lost one sixth of the progress we’d already made. We expect a modest decrease of 22 million people to fall out of poverty in 2021, setting the world back 6 years in terms of reductions.
Even without the alarming disparities in vaccination rates (that were wholly predictable), slowing growth in many of the countries affected (such as Brazil) will make recovering lost ground a tough task. Additionally, access to running water, electricity, food, education, and healthcare declined as recently urbanized nations saw their population run back to the countryside to avoid the virus. Insufficient government stimulus and lack of international aid, plus the sharp drop in employment, will have permanent consequences - but the direst effects will be certainly in education. The Netherlands, one of the richest and most well connected countries in the planet, reported what amounted to a literal lost year in terms of learning via online school.
Manna from heaven
Fortunately, there is one good cure for poverty: money. Regardless of how you define it, poverty always ends up being “not having enough money”. While 1.9 dollars is the escape from absolute poverty, 5 dollars is (roughly) the average poverty line, and 10 dollars is the daily consumption associated with a permanent escape in poverty. While income and wellbeing are relative (for example, Argentinian income grew in 1994 to 1998, but welfare actually declined) it is needless to say that more money is more good. The key question is: does economic growth or redistribution matter more?
Although I don’t believe that there is a trade-off (and nobody’s claiming there is), this is a victory for team redistribution: an increase in GDP of 1% reduces poverty by roughly 3%, but a reduction of 1% in the Gini coefficient within each country reduces it by about 6%. This is, however, sensitive to two measurements: the poverty line, and the degree of poverty at any given line. These sensitivities occur because transferring at high poverty lines (say, 5.5 dollars a day) to the poor from the rich, in very poor countries, can push the people right above the thresshold of poverty back below the poverty line. This is unusual, but it does suggest a real tradeoff between inequality and growth (in the opposite direction as assumed), and/or that there is a limit to pure redistribution of income.
Now, I don’t think you can get from extreme poverty to no poverty by pure redistribution. The main reason is clearly empirical: the vast majority of reductions in poverty rates have come from “middle of the road” countries becoming dramatically less poor, and not really from uniform movements. To bring back a previous example, East Asia slashed poverty from 58% to 1.5% - whereas Subsaharan Africa only cut it by 25 points, to the mid 30s.
The second reason is more practical: redistribution costs money, not just in terms of bureaucracy, but also in terms of actually having any money to redistribute. As stated above, trying to spread around a very thin cake can be counterproductive in some very marginal cases. Just as Amartya Sen’s study of famines found that the main reason that people died wasn’t that there wasn’t enough food for them, but rather that they couldn’t lay a claim on any of it, recent research has shown that the national governments of almost all countries have access to enough money to redistribute it towards erradicating poverty.
What about jobs? Isn’t having economies grow and create jobs the solution to poverty? Well, yes, kind of. Adults with formal jobs (informality is a huge issue in the developing world) are rarely ever poor. The real problem is that most poor people are usually elderly, children, or non-working women: people who can’t actually “just get jobs”. And even if other people get jobs and provide for them, it’s not clear that they can get jobs that provide enough for a whole family. Based on OECD data, increasing employment would be more difficult than just giving children money, for any given rate of child poverty (the initial rate is really important).
What can Africa tell us about fighting poverty?
Paul Krugman once said that economists understand something when they can, in the terms of Thomas Schelling, tie micromotives (i.e. individual actions) to macrobehaviors (i.e. large scale phenomena). So just like we can’t know whether a butterfly causes a hurricane, can we know whether giving people money to end poverty will have horrifying, counterproductive results? We can, and the answer is no.
The first kind of evidence comes from Randomized Control Trials (RCTs). RCTs are a type of experiment where, of a group of subjects, half gets something and the other half, called the control, gets nothing. For instance, you might give a bag of lentils to people who get vaccinated in half of the villages in a region to test whether it makes them more likely to get shots (Nobel Laureates Abhijit Banerjee and Esther Duflo actually tried it, and it works).
A series of projects in Africa, centering on the charity GiveDirectly, wondered: does giving people money have bad consequences? After doing a trial, it turns out it doesn’t, consistent with other experiments in similar developing countries. Giving people cash is also better than other usually proposed programs, like job training (per the Kenyan RCT), food distribution, etc. Another approach that might work is called “poverty graduation”, and generally consists of both giving people cold hard money, valuable assets (like cows, or bicycles), and mentoring. In some cases, though, graduation costs much more than cash for the same advantages.
What can Latin America tell us about fighting poverty?
The other useful case of how giving people money can (not) help them comes from Latin America, via what are known as conditional cash transfers. Conditional Cash Transfers (CCTs or cash transfers, from now on) are, as the name says, cash transfers with strings attached: you get a monthly payment if and only if you fulfill certain criteria, generally related to enrolling children in school and keeping their vaccinations in check.
The best regarded of these is Brazil’s Bolsa Familia, which was both highly successful at getting people to sign up, at decreasing inequality, and at reducing poverty. Similar Brazilian programs, such as Bolsa Escola, have been less successful at reducing the poverty headcount but effective in getting people to go along with the conditions. Studies of Honduras’s PRAF show that the poorest people (bottom 20%) were the most affected, that CCTs were effective at promoting enrollment in school, and that there were no negative effects on labor supply. Studies of Argentina’s AUH, which have found much more moderate success (perhaps because of the scope of the program), have actually found that it discourages the transition from informal to formal work, which might be a big problem given the prevalence of labor informality in the region.
In fact, Latin America has been wildly successful at decreasing income inequality in the past 20 or so years, partly because of a drop in the premium to higher education (which is a whole thing, but it’s not particularly good) and partly because of big redistribution programs. The 2000s were that rare time where poverty and inequality declined while income grew - the 90s growth phase had a much smaller decline in poverty, and an increase in inequality. Other policies, such as higher minimum wages, have also had a positive impact. However, items such as a big gap in access to education (particularly, paradoxically, higher education), healthcare, and new technologies.
What can rich countries tell us about fighting poverty?
Rich countries, by virtue of having high incomes, have very low poverty rates (by global standards). As seen in the chart above, if reducing inequality is a component of the reduction of poverty (as well as growth), then developed nations are generally very good at doing that - or can be. Only South Korea (which is fairly equitable) and the not actually developed nations (Chile, Mexico, and Turkey) don’t have large reductions in inequality from their taxation and spending.
If you look at the World Bank figures for Europe and North America, extreme poverty barely increased last year - and American stimulus was actually so generous that the poverty rate decreased to 8.5%, the lowest figure ever recorded. Using a different measurement, it was halved compared to 2018. Child poverty was slashed from 13.7% to 5.9%, rather than the alternative: a sharply higher interest rate, at least as high as during the 1970s. The US actually spent as much on stimulus as it did during 1943, at the height of World War Two, and it had the largest stimulus package of the developed world, followed closely by Singapore and Slovenia. The disposable income of Americans grew in the pandemic, versus the decreases of the Europeans.
Speaking of the US, it is a bit unfairly maligned in terms of poverty. It has a much smaller welfare state than Europe (and also a much more expensive healthcare system, which hides how small it actually is) but it has accomplished quite a bit. Contrary to popular belief, Lyndon Johnson’s War on Poverty was actually a huge success, and the key is pretty mundane: poverty is measured, to be blunt, badly, with a rate that adjusts for things like cost of living or changes in consumption patterns.
The official poverty rate in 1964, when the War on Poverty began, was a striking 19.5%; whilst the official rate is 10.5%, the rate after adjusting for bad methodology is an astounding 1.6%. And using current standards, the poverty rate is an unimpressive 10.5%, although it’s worth pointing out that nearly 70% of Americans were poor in 1963 under this measure. Obviously this means that the US should set better poverty thresholds, but it also means that it can actually, effectively attack poverty if it wants to.
Growth still matters
Lastly, why does economic growth still matter if redistribution can get you all the way?
Several points that have been made are tempered by growth. For instance, the issue of resources: while developing countries do now have the resources for redistribution, they didn’t in the past, and they might not have the capacity to implement it (or the institutions to want to do it) for a long time. Plus, the amount of pre-redistribution poverty (aka “market poverty”) matters, because countries with more income need less money to eradicate poverty, and thus have more money for anything else they might want. And rich countries were only able to afford such large COVID stimulus packages because of their wealth - there appears to be a relationship between GDP per capita and size of the stimulus: Argentina only spent 3% of GDP, whereas Australia (5 times as rich) spent 13%.
There’s also the fact that inequality reductions might have decreasing returns: countries with high poverty and/or high inequality reduce poverty twice as much as other countries through redistribution, and the amount of growth that benefits the very poor is basically zero. But in less unequal countries, the “elasticity of poverty to growth” is basically one, meaning that the rising tide does lift all boats. It seems that very few countries actually have growth benefit the poor specifically a lot, and most countries can have strong transmission between growth of average income and growth of lower incomes past a certain point in equitability.
Plus, reducing poverty is also a pro growth policy. Since poor people have less human capital accumulation, reducing poverty increases the human capital stock (this is why Latin America getting those inequality gains by cutting the university wage premium is such a dicey bet, too). As a result, tackling poverty (and other related factors of underdevelopment) results in a stronger base for the economy, with more broadly shared growth.
Even if you don’t think economic growth is necessary, reducing inequality is enough to create it - which is particularly important given that the efforts of getting more and more improvements in poverty with little growth are really hard. This thread by economist Leopoldo Tornarolli, for instance, shows that halving Argentina’s poverty rate to 21% would require slashing inequality by 45%, turning the country into one of the world’s most equitable. And the costs of reducing inequality that much would be staggering as well, which would probably be very high to cover without the country actually getting any richer.
In conclusion, extreme poverty is bad. With minimum effort, we’d be able to erradicate the worst kind of poverty in the world within our lifetimes, maybe more. I am not some insane goalpost shifter, but after having everyone cross the 1.90 threshhold, we’d focus on crossing the 2.5 line, or the 3.2 line, the 5.5 line, the 7.4 line, the ten dollar line, etc. Just like the US’s War on Poverty was successful and now needs a do-over for new times, we should do the same for global poverty, based on redistribution to the poor, economic stability, and broadly shared growth.
Progress before and until COVID
Just give people money
Africa and poverty
Latin America and poverty
The developed world and poverty
Why growth still matters