Can Better Job Offers Fight Modern Slavery?

In my last post, I mentiond that James C. Scott has this concept called legibility. Essentially, institutions need to abstract complicated social realities into discrete categories they can analyze and act upon. An institution can’t remember “Micah, who is sort of like this in this context but actually completely different when he is around these people.” It needs: Micah Zarin, age 18, student, income Y. After all, the word abstract comes from the same Latin root as subtract: trahere, meaning to draw or pull away. To abstract is to pull away, or to subtract certain details. According to the 2022 Global Estimates, there are roughly 49.6 million people in modern slavery on any given day. That's one in 150 people alive. And the illegal profits from forced labor are estimated at $236 billion a year, about $10,000 stolen from each victim, every year. A UN University project analyzed over two million aid records and found that, on average, less than $12 per victim per year was committed to fighting modern slavery. The ratio of money made doing the harm to money spent stopping it is roughly 800 to 1. When I was searching through the EA Forum about this, I found a thread from December 2023 where someone asked why modern slavery isn’t a bigger cause area. @Karthik Tadepalli basically said that nobody had really thought about it, while a Founders Pledge researcher, @Rosie_Bettle , noted that the evidence base for direct work is surprisingly weak. I'm going to talk about kilns for the purposes of this post, and I will offer a proposal that I hope can be implemented. If you are a legal scholar, a development economist, or someone with field experience in South Asian labor rights who thinks this might be a good idea, or has reasons this migh be a bad idea, please reach out to me. I'd love for this to turn into a project. I used AI to help find resources and data on these issues. All data I fact checked myself I think Rosie is right. EA relies heavily on data, and it is very difficult to get data about slavery. For this post, we will focus on brick kiln owners in South Asia. A commercial brick kiln is an open air industrial complex that is used for manufacturing bricks. It's extremely labor intensive and relies heavily on manual work. Entire families, including children, dig clay by hand, carry the clay in heavy baskets on their heads, mold the clay to bricks, and then stack the bricks to be baked. After baking, they break the bricks apart and carry them again. This runs up to 14 to 16 hours a day. In an ILO assessment of Afghan kilns, they found that both children and adults work over 70 hours to per week. They live in flimsy, windowless mud huts built directly on killing grounds, and the owner has almost complete control over their lives. The workers live on the grounds, working 70+ hours a week. They have no private time, no private space, and often no access to phones or outside communication. There is no "back channel" for information to leak out. Unlike a factory in a city where workers might go home at night and talk to neighbors, these families are isolated in a remote industrial zone. They control the shop where the workers buy food on credit, which the ILO identifies as a major driver of dependency. And in many cases, the owner or his relatives have influence over local elected bodies and the police, which the same ILO paper lists as a reason that workers are unable to take legal action. The primary mechanism of control in this system is actually not the kiln owner but rather the labor contractor, known as the naike or sardar. The naike is basically the person who recruits workers from their home villages. A study by the International Labour Organization and UNICEF in Nepal found that 75.7 percent of kiln workers receive their cash advance from the naike, while only 3.5 percent receive it directly from the kiln owner. The naike travels to the home villages of potential workers before the kiln season begins, which usually runs from November to June. He offers the family a cash advance, typically between 250 and 600 US dollars, which is used by the family to cover lean-season expenses, medical emergencies, or social obligations like weddings. And the family is essentially forced to accept the advance, as they have no other source of credit. After all, banks don't lend to them, since they have no collateral, and moneylenders charge interest rates as high as 10 percent per month. The naike, on the other hand, offers the money with no explicit interest. But, in exchange, the family owes him a season of labor. Now, the terms of this agreement are rarely written down. Instead, the debt is recorded in the naike’s personal ledger, which the workers cannot see. Then when the season ends, the naike presents the family with a settlement that often shows they still owe money, due to deductions for food, tools, and alleged errors. This debt rolls over into the next year, binding the family to the naike indefinitely. Importantly (for reasons you will see soon), the naike is usually not an outsider. He (usually male) is from the same village, often the same caste, sometimes a relative, and a lot of the time he started out doing the same work. So, to the villagers he actually looks like somebody who made it. But his whole income is a cut, in Pakistan it is 20 bricks out of every 1,000 a family makes, for as long as that family stays at that kiln, and he is often financed by the owner himself, so when a family takes his money and doesn’t show up, that debt is his. He tells the families is that he is the only one who will lend them money without interest, and that he is keeping them fed until the season starts. Uncomfortably, researchers in Tamil Nadu found that the workers themselves favor the advance system, out of necessity, poverty and fear, and a study of Chennai kilns found workers and owners both endorse it and the workers see it as their one shot at moving up. Starting in June 2000, the ILO ran a Dutch-funded project across Bangladesh, India, Nepal and Pakistan that piloted microfinance-based schemes to prevent debt bondage and to rehabilitate people released from it. Two people in its Social Finance Programme, Patrick Daru and Craig Churchill, wrote up what they had learned in the first eighteen months. In their conclusion, they noted that where labour is plentiful and jobs are scarce, freeing every bonded labourer without access to alternative employment is not a solution for the families. Also, employers would just respond by hiring other migrants on the same terms, or by automating jobs. What they proposed instead was to study the labour and credit markets from the employer's side and then invite a handful of employers into what they called "small demonstration projects to try out alternatives to bonded labour arrangements." The idea was that the workers would keep their jobs on better terms, and other employers would get to see, at a kiln down the road, that a substitute for the debt existed and worked. They were careful to add that this only works alongside a credible threat of prosecution for the worst cases, and that treating employers as criminals would probably get less done than talking to them. This is because up to a certain level of debt, the arrangement works for the employer because the family will stay through the peak season, and it works for the family because they get a guaranteed job, an advance they may never fully have to repay, and an informal safety net if someone gets sick. Only past that level, when the debt grows so large that everyone knows it will never be repaid, does the employer switch to violence and restrictions on movement. As far as I can tell, the ILO mostly never ran the experiment it had just proposed. The project eventually grew into PEBLISA, a larger program funded by the Netherlands at around $2.4 million, which did lead to various useful outcomes. In India, for example, the work fed into a review of the Bonded Labour System Abolition Act and later into a “convergence approach” that tried to connect bonded laborers with existing welfare schemes. The closest thing I could find is Better Brick Nepal, which began in 2013 with five kilns. It was initially supported by Humanity United, and eventually grew to around forty partner kilns out of roughly 1,200 registered kilns in Nepal. Fifteen kilns were certified between 2013 and 2021 before the program was handed over to a more locally led initiative. What I find interesting about Better Brick Nepal is that if you read the Better Brick Nepal standard, it does not say that advances are inherently abusive. In fact, it says fairly explicitly that an advance against future wages is not, by itself, bonded labor. Instead, it tries to make the arrangement more ethical. For example, the kiln should show that the advance can realistically be repaid within one season. Furthermore, there should be a written contract, explicitly detailing the terms of the work. The standard even asks for a retaining allowance for workers who come back for multiple seasons, essentially paying them part of their remuneration during the off-season. To be clear, these are progress criteria rather than requirements. So, for example, the standard asks kilns to work toward things like more regular and individual wage payments and retaining allowances during the off-season. This matters a little, but not really for the point I am making. Better Brick Nepal still establishes something that, before it existed, we would have had to speculate about: a commercial kiln can operate while moving away from debt bondage, without everyone immediately saying that it is economically impossible. A bank theoretically sells money, except these families often cannot borrow from a bank. A microfinance institution can sometimes sell money, but that doesn't guarantee a job at the other end of the loan. A government employment program can offer work, but not necessarily at the time the family needs it, and historically not necessarily with reliable payment. The naike gives you the advance before the season, gets you and usually the rest of your family to the kiln. And if something goes wrong halfway through the year, there is a person who can extend the ledger again. The fact that the ledger is also how the family becomes trapped is exactly what makes this so difficult. The ILO basically noticed this twenty-five years ago. That is why Daru and Churchill were so worried about simply "freeing" people without replacing what the bonded arrangement was providing. Nepal had already given them a natural demonstration with the Kamaiya system: if you legally erase the debt while leaving people with no land, no job and no functioning safety net, the old social relationship will simply reconstruct itself. Some former Kamaiyas went back to repaying debts that the law said no longer existed. A few years ago, Karthik Muralidharan, Paul Niehaus and Sandip Sukhtankar studied a reform to India's rural employment guarantee. The reform basically made the program function better: payments became more reliable and the government job became a more credible thing a worker could actually expect to get paid for. Because the reform was randomized across large subdistricts covering roughly nineteen million people, the researchers could look not only at the workers directly using the program but at what happened to the local labor market around them. The results are kind of insane. Beneficiary household earnings rose by 14 percent and poverty fell by 26 percent. But only a small share of the income gain came directly from the government jobs. Eighty-six percent came from outside the program, largely because private-sector wages and employment increased. This is because in a perfectly competitive labor market, workers can easily leave one employer for another, so employers have very little room to push wages below what workers are worth. But rural labor markets often do not work like that. Workers may have only a few realistic employers, little information about other jobs, limited ability to travel, and an urgent need for income. That gives employers some power to set wages. The reform weakened that power by improving workers’ outside option. Before, threatening to leave a private employer was not very meaningful if the alternative government job was unreliable or might not pay. Once the government job became credible, workers could actually say no. Their reservation wage (the lowest wage they were willing to accept from a private employer) went up. And then private employers then had to raise wages or otherwise make their jobs more attractive to keep workers. I think the same principle could be tested in brick-kiln recruitment. Instead of simply telling families not to take the naike's advance, we could go to sending villages during recruitment season and create a second complete offer with more ethical conditions. Once that competitor exists, the other kilns have to respond. They can improve their terms, and if they do that, the program has succeeded even among families who never switch. Or, of course, they can just intimidate workers. So, the program should have a lawyer and a relationship with the district labor officer before anything. If you are a legal scholar, development economist, funder, or someone with field experience in South Asian labor markets who could help stress-test or run this experiment, please reach out.