Was banning Indian child labor counterproductive?

The authors are Prashant Bharadwaj, Leah K. Lakdawala, and Nicholas Li, and here is the abstract:

While bans against child labor are a common policy tool, there is very little empirical evidence validating their effectiveness. In this paper, we examine the consequences of India’s landmark legislation against child labor, the Child Labor (Prohibition and Regulation) Act of 1986. Using data from employment surveys conducted before and after the ban, and using age restrictions that determined who the ban applied to, we show that child wages decrease and child labor increases after the ban. These results are consistent with a theoretical model building on the seminal work of Basu and Van (1998) and Basu (2005), where families use child labor to reach subsistence constraints and where child wages decrease in response to bans, leading poor families to utilize more child labor. The increase in child labor comes at the expense of reduced school enrollment. We also examine the effects of the ban at the household level. Using linked consumption and expenditure data, we find that along various margins of household expenditure, consumption, calorie intake and asset holdings, households are worse off after the ban.

I’m not trying to talk you into child labor with this post.  Rather, you should be less confident in a lot of your moralizing about what is a good policy or an evil policy.

Hat tip goes to Dev Patel.


so why did it (appear to) work elsewhere?

Perhaps the alleviation of subsistence-level poverty? Is there somewhere that has a significant amount of people struggling to reach subsistence level, that has successfully eliminated child labor?

China under the Communists? Just to throw it out, without attempting to define what is a good polity or an evil polity.

While I'm not an expert on what is happening on the ground in China, many sources are claiming that child labor is rampant there.

If you mean a previous era, child labor was also used under Mao in the Great Leap Forward.

Just looking very quickly, it may hinge on definition. For me, child labor means child is defined as someone under the age of 13 or so. A working 14 year old would not be considered the same thing as a child, but would fall under the category of 'youth.' (As might a 12 year old, though not a 10 year old, actually.)

China was a mess under Mao, but the Communists did seem taken with the idea of indoctrination/education. Say what one wishes about China, but the notable increase in literacy was likely coupled with a notable decrease in child labor. Though this is the MR comment section, so it likely that at least some commenters would consider teaching children how to read the equivalent of forced labor.

I'm guessing it was a post hoc ergo propter hoc kind of situation. Sort of like OSHA: http://www.themoneyillusion.com/wp-content/uploads/2011/03/osha.png

"so why did it (appear to) work elsewhere?"

Maybe it didn't?

Child labor law enforcement was stricter in the 60s than enforcement seems to be today. Real wages had to be higher.

Child employment then was higher then than today in the US.

1. I found the proposed explanation of this thread interesting:


2. Aside: one should also ask the question of whether the spread of education has been a net harm for India: given the atrocious quality of Government schools in rural areas, it is not, at least a priori, clear that a lot of the children couldn't have used their time more fruitfully than by going to school.

"Children must get education" may be a moral claim where actual results go unexamined (much like banning child labor).

'I’m not trying to talk you into child labor with this post.'

So coy.

Silly and navel-gazing. Of course banning child labor is the morally correct thing to do. If in this case child labor somehow increased after the ban, as suggested, then that only means that the ban was improperly implemented or enforced. But the idea behind the ban is self-evidently a good idea.

Far from self evident. Some of us distinguish good intentions from actual effects in evaluating policy, which is important because very often each leads to different conclusions.

If this regulation worsened child labour rates then there is nothing to admire about it.

'If this regulation worsened child labour rates then there is nothing to admire about it.'

Because clearly it failed in its purpose, not because its purpose was wrong- Regardless of how coyly Prof. Cowen would like to suggest that policy based on 'moralizing' needs to be evaluated not by its success or failure per se, but on the fact that moralizing is involved at all.

Unlike the moral clarity provided by clarion calls for increased economic freedom, of course.

So your conclusion from this is that Tyler thinks we should not evaluate policy based on evidence?

Oh no, far from it. We should avoid letting any moral principles outweigh policy failure when attempting to take those small steps to a much better world, and thus change policy to make banning child labor, on moral grounds, more effective.

Banning child labor is clearly a moral position, and equally clearly, any attempt to ban it must be weighed against a particular policy's failure to do so. In other words, failure must clearly mean that morality is a poor way to create policy, right?

I would say that morality is a good motive for policy, but a poor substitute for evidence of effectiveness.

Pretty much, but remember that Prof. Cowen is strictly opposed to anything like virtue signalling. For example, saying that everyone in a society should have access to health care is clearly virtue signalling, and thus prima facie not a good basis for formulating policy.

It's not silly.

Child labor is bad. Starvation is even worse. Some people may have to choose between the two.

Child labor is bad. But in a given governance reality, it might not be possible to implement a ban that effectively reduces child labor, and in fact in some cases a ban could even make things worse.

It is indeed self-evident that child labor is not a good thing, is immoral, and is something we hope to see eliminated. But you can't therefore ignore the consequences of actual policy. The goal is noble, but then whether you are achieving the goal matters all the more (not less).

I'm a bit confused about why child labor is "self-evidently" bad? Is a paper route bad? Do people really mean "all children should get an education"?

I was thinking the same thing. I mowed neighbours' lawns from ten years old, babysat from twelve years old, all of which now seem to me to be good things. I guess if work were all I were doing (and not going to school). it wouldn't be so good? Is that the moral of the story? Or maybe the age the labour takes place?

Isn't the solution then to combine child labor bans with subsistence level UBI, rather tan throw up ones hand and just accept child labor as inevitable?

Given limited political capital and resources, the best solution is "What few initiatives can we prioritize that are most likely to do some good and put us on a path to success?"

Eliminating child labor should be an eventual goal, but banning it may not be the best short-term policy. Choosing not to immediately ban it is not the same as accepting it as inevitable.

Banning child labour imposes naive western ideals on complex problems...

So, child labour is a Giffen good for poorer Indian households?

"I’m not trying to talk you into child labor with this post"

But... you're not exactly trying to talk us out of it either.

child labor shoud be only be banned in degrading work, like mining. no wonder singing and acting is not banned.

You have a point here. What, exactly, are we talking about when we want to "ban child labor"?

Child entertainers face similar downsides to children doing more mundane labor. They talk of being deprived of their childhood. But this is legal and ok, apparently. Of course, if they become a star, there are benefits that (maybe) outweigh the negatives. But there are surely lots of non-star child entertainers for each one that becomes a star.

How does the cost-benefit of labor for a child in a rich western country potentially becoming rich and famous compare to the cost-benefit of child labor in the context of extreme poverty?

Fair point. The degrading, or dangerous work is what I think most want stopped. No one probably wants to stop chores or household work that gets used in theory toward their own good (farm labor probably fits in here). One also hopes schooling fits in to 'not work', but to the 7 year old, I'm sure that seems like work- being stuck in a chair and not playing.

Banning only degrading work doesn't change anything.

Dessy, S. E., & Pallage, S. (2005). A THEORY OF THE WORST FORMS OF CHILD LABOUR. The Economic Journal, 115(500), 68-87

"Although intuitive and morally compelling, a ban on the worst forms of child labour in poor countries is unlikely to be welfare improving. We show that harmful forms of child labour have an economic role: by maintaining wages for child labour high enough, they allow human capital accumulation in poor countries. Unless appropriate mechanisms are designed to mitigate the decline in child labour wages caused by reduced employment options for children, a ban on harmful forms of child labour will likely prove undesirable. We perform our analysis within a simple model of parental investment in children's education."

23 comments and no one has arrived to this part of the article:

"Using a more refined difference in difference approach dictated by the model, our results show that a child between the ages of 10-13 with a sibling below the age of 14 significantly increases his or her labor force participation by 0.8 percentage points compared to a child of the same age with a sibling over the age of 14, which is approximately 5.6% over the pre-ban average participation for that age group".

It seems the researchers discovered elder siblings provide for the younger. So, does this familial investment strategy (elder support the younger) pays in the long term?

I've referred to this paper elsewhere and was stoutly told off in the comments that the paper never went any further. That there was some error with it. Not sure if this is true or not. But I certainly got shouted at for using it.

As to why child labour (might) increase with a ban, the child wages are what stands between the household and starvation. If child labour is banned then wage rates for it fall, meaning more hours of it must be supplied to keep the household from starvation. This is thus something that applies at subsistence levels only....

Dev Patel has let himself go since Slumdog Millionaire.

Even in the U.S., children work in farma and others businesses of their parents or close people. That is often good for their children and should not be eliminated. Only harmful child labor should be eliminated.

Anon likely has the correct call on this one (and many other studies cited on this blog): post hoc, ergo propter hoc. In any case, urbanization may be the most significant cause of the increase in child labor: https://blog.timetoswipe.com/2016-report-child-labour-in-india/

Thankfully, Tim Taylor offers guidance (and references to Hayek, Stiglitz, and F. Scott Fitzgerald): http://conversableeconomist.blogspot.com/2017/09/imperfect-information-is-it-for-or.html

So Tyler feels child labor is an important contributor to family welfare, and should not have been banned. Or as a compromise, would he suggest that the ban be revised to permit families to send only the eldest child off to make rugs, as a sacrifice for the others?

I've noticed this is not something Americans choose to care about very much, judging from their fondness for those very pretty, inexpensive hand-knotted or hand-tufted rugs from India. I believe, though, that people might have compunctions about buying Indian rugs if the tag stitched on the bottom said "Proudly Made in India with Child Labor."

If instead of a ban, such a tag had caused rug sales to decline, I wonder if Tyler would then suggest that the truth about child labor ought to be concealed?

Tyler's post is revealing a truth, and you're the one struggling with it. Suppose this truth had been known from the outset, and the bill banning child labor contained a statement acknowledging, "This bill will increase child labor." Would you have counseled your MP to vote for it?

It is an unusually murky "truth," if so - there being no evidence that the ban caused the increase in child labor, and didn't prevent it from being even greater.

If you pass a law banning X, and X increases after that, then at a minimum your law has utterly failed. Nothing murky about that.

Establishing that the ban caused the increase is harder, but the evidence they provide is pretty compelling. There is a plausible mechanism for the ban causing an increase, and the data matches the expectations for that mechanism in several ways (e.g. wages going down).

No, not at all murky, especially if the addition of half a billion people since 1986 is a fact that bores you.

I would not automatically assume that the intended purpose of child labor laws is to protect children. In the US the first federal child labor law was introduced in order to restrict competition. It was lobbied for by businesses in states that already had state laws prohibiting child labor, on the grounds that firms in states that lacked such laws had unfair access to cheaper labor.

Well...ok, but presumably those state laws were passed with the intent of protecting children. And the fact that business lobbied for it doesn't mean that wasn't the only reason the law got support. The federal law probably had enough support to pass because there were also people who wanted to protect children that supported it.

Some people, like myself ,cannot "put aside" their morals to entertain whether something is market efficient or not. The "market" is not everyone's "be-all-end-all" .

Can you "put aside" your morals to entertain whether you are helping or hurting people?

My morals, as instilled in me by my Catholic upbringing, do not entail hurting anyone. In fact, the are explicit in guarding against just that.

Perhaps you can elaborate?

Your morals may explicitly guard against hurting someone, but given that they don't harm doesn't mean they help. Your initial post implies that banning child labour is the morally correct position. What's the point of banning child labour if the ban doesn't improve the welfare of the family?

Market efficiency may not be the "be-all-end-all" but if a ban on child labour results in an increase cost for a family - perhaps employing the next oldest child in addition to the oldest, or the search cost of replacing the income lost by the ban, or just leaving them with less income than before -- isn't the ban failing your morals and actually hurting someone?

“Your morals may explicitly guard against hurting someone, but given that they don’t harm doesn’t mean they help.”
Correct, but you do not define what you mean by "help" My morals do not always help. They are not meant to. I would have been an abolitionist back in the day. My morals would have done great economic harm to certain plantation owners – that’s for sure.

“What’s the point of banning child labour if the ban doesn’t improve the welfare of the family?”
You define “welfare” in monetary terms, as you mention “an increased cost” in your second paragraph. This increase in cost is, in your opinion, is worth avoiding by having young children work.
What about the cost of having children work? I guess it is because they are children and of not much economic significance that you do not consider the benefits they reap by NOT working -i.e. going to school or just being children not burdened with job. If this is a solution, in your opinion, in this particular case, why not re-institute it here in the States for those poor families at the bottom of the economic rungs?

What about the apparent fact that at least in this case banning child labor actually resulted in more child labor? Does that matter? Or is signaling your virtue by supporting a ban more important than the additional children forced into the labor forever by the ban?

Generally, the bulk of child labor is children working for their parents on a farm or family business.

We have long considered that a different form of child labor that does not carry the same negative consequences. I picked cotton as a first grader in 1947 and did not consider it a bad thing. Actually,
in the southern rural community the school system divided the summer break into a spring and a fall break with school resuming in the middle of the summer so children could help with planting and harvesting and not miss school.

I do not know about India, but I have to wonder how much of the child labor in India is kids working in a family business or farm and this may not be as adverse a results as the paper implies or assumes.

P.S. Everyone, adults and children was paid piece-rate (pounds of cotton) so minimum wage was not a factor.

The right question regarding child labor is not "should it be legal or not" but rather "at what point in a country's development does it become unnecessary for parents to employ their children as workers and instead put them inside classrooms?" Ban child labor too early and you end up with children on the street seeking prostitution or begging for money. Yet ban child labor too late and you lose out on future generations of economic growth because when you should have children learning in school, they're instead wasting their human capital on working in hazardous conditions.

As a comparison, child labor in the US was not fully banned on a federal scale until the 1938 Fair Labor Standards Act.

US GDP per capita at the time in chained 2009 dollars was about $7692 ($1 trillion / 130 million, note Angus Maddison has it more like $6126). India nominal GDP per capita in 2016 is about $1709. For some reason, India GDP per capita PPP is around $6500, but I'll let you decide what is more developed, the US in 1938 or India in 2016.

Trainning their children to fight for their rights or die trying would prolly help families to reach subsistance levels faster : if we're looking into policies without morals, far more options than the basic submission to lanlords are open.

Often it is best for the children to be at school for some hours and at work for some hours. However, Indian public schools are sometimes useless.

Child labor is considered morally unacceptable in many areas of the world. India enacted the Child Labor (Prohibition and Regulation) Act of 1986. A paper written by Prashant Bharadwaj, Leah K. Lakdawala, and Nicholas Li tries to answer the question, “Was banning child labor counterproductive?” While child labor is completely unethical, it has economic implications in industries in third world countries. Instead of looking at child labor as a moral situation, the researchers looked at it as an economic situation. They analyzed the costs and the impacts of wages that the child labor ban had on Indian households. I agree with the researchers that the child labor ban was counterproductive on an economic level.
Looking at the demand side of products the family could buy, the child labor ban was counterproductive. One of the determinants of demand is buyer expectations, and among these is future income. The wages that were given to a child for their labor was income for the family, and this income would have increased their disposable income. By increasing their disposable income via wages from child labor, demand of products is increased since families now have more buying power. However, with the child labor ban, families no longer have a large amount of disposable income, causing the demand for normal goods to decrease and the demand for inferior goods to increase. Another factor of the child labor ban is opportunity cost. In the instance of child labor, the next best option for the child would be to work on a family farm, if their family owned one. However, on a family farm, since the child is considered part of the family, the child usually receiving no wages, and does not improve the family’s income.

Comments for this post are closed