Private Schools in Developing Countries

Tina Rosenberg has an excellent piece on private schooling in developing countries at the NYTimes blog:

In the United States, private school is generally a privilege of the rich. But in poorer nations, particularly in Africa and South Asia, families of all social classes send their children to private school….

BRAC used to be an acronym for Bangladesh Rural Advancement Committee, but now the letters stand alone. It was founded in 1972 to provide relief after Bangladesh’s war of liberation. Although you’ve probably never heard of it, BRAC is the largest nongovernmental organization in the world, with some 100,000 employees, and it services reach 110 million people.

…And since 1985, it has run schools… BRAC has more than 1.25 million children in its schools in Bangladesh and six other countries, and it is expanding.

BRAC students, in fact, do better than their public-school counterparts….BRAC students are more likely to complete fifth grade — in 2004, 94 percent did, as opposed to 67 percent of public school students. (The BRAC number is now about 99 percent.)  On government tests, BRAC students do about 10 percent better than public school students  — impressive, given that their population is the most marginalized. (emphasis added).

In my own work on private schools in India I also found suggestive evidence that private schools–mostly very small, urban slum schools–produced better outcomes than their public counterparts (paper (pdf), video).


Another point to note – and I don't know how true this is across all developing nations, but for India at least – is that private school teachers (outside of the middle-class and elite) are paid substantially less than public ones.

I know in parts of India the difference is very stark (leading to creative and corrupt arbitrageurs). The general success of private schools versus the state is an important check to the common claim an easy way to improve outcomes is to pay teachers better unconditionally (as a means of incentive there are perhaps more complex outcomes – and I'm sure there are a 100 RCTs studying one thing or the other wrt this).

It is invalid to make any private-versus-public judgement from these studies at least in the Indian context. They, almost by design, cater to a different audience (though the government may not admit it). The strictly public schools cater to parents who could't afford a private school, sort of a public net to prevent students falling through the cracks. Or in areas so remote etc. that no private entrepreneur wants to run a school.

No doubt they are worse, but a part of that is simple because they are catering to the lowest common denominator and accepting the toughest students. Even the relatively poor parents send kids to private schools if they can. This may be a feature and not a bug.

I'm no apologist for the flaws and corruption of the Indian public school system but using that example to make a public-private quality comparison is misleading.

I was thinking the same thing. How much of these results would survive a randomized trial?

You're right about all the caveats, but I think this criticism as actually much more appropriate in an American context. In India, there are many private schools (run usually by missionaries, NGOs, or charities) that cater to the same *or worse* demographic as the government schools. In the US, private school kids are almost exclusively richer. And so people can run an experiment for kids who otherwise would have been in a public school.

This is an important distinction, no doubt, but do note that BRAC students in the article were said to be more marginalized.

When, as in the United States, 10% of the kids go to private schools it makes sense to wonder whether the 10% are markedly different than the rest, the cream. But in some regions in India more than half the students are in private schools which makes the cream skimming argument less plausible. See my paper or the video for some evidence.

Nice paper. Is there a way to access and play with your dataset?

Rahul, the primary data set is here and includes lots more things of interest

Yep, this was my point exactly.

And as your paper notes, the variance in district wrt private schooling offers good reason to believe there isn't strong selection biases etc.

You can go about creaming in education either by over sampling the easiest children to educate or by under-sampling the hardest children to educate.

A sample that contains 50% of the total population might not easily represent the former, but could still easily represent the latter.

@Mike Hess:

Yep. You can improve metrics by skimming the top or by avoiding scooping up the dregs.

OTOH, Alex's paper does have some reasoning as to why that's not happening, but the data doesn't seem very convincing.

I would not generalize from one observation but private schools in the developing world come in 2 forms: Schools founded with academic success in mind that hire experienced teachers and are highly selective regarding the composition of their student body, and schools founded with profit in mind that serve as release valves for students that cannot make it in public schools. I'd love to see a serious study including both before I can make confident statements regarding private and public schools. My guess is unless unless you're already a bright and determined student, private school won't make a difference. This has actually already been looked at here.,8599,1670063,00.html

For a more wide ranging review of the private schools for the very poor in India and Africa you want to read the book "The Beautiful Tree" by James
Tooley. Lots of good information on both government run schools and private schools paid for by concerned parents. Good read.

Then there are the countries like Germany, where the term private school is probably not what one would customarily use, even if the schools are 'private' (a right ensured by the Grundgesetz). And where the results are essentially difficult to distinguish from 'public' schooling -

'Ersatzschulen are ordinary primary or secondary schools, which are run by private individuals, private organizations or religious groups. These schools offer the same types of diplomas as public schools. Ersatzschulen lack the freedom to operate completely outside of government regulation. Teachers at Ersatzschulen must have at least the same education and at least the same wages as teachers at public schools, an Ersatzschule must have at least the same academic standards as a public school and Article 7, Paragraph 4 of the Grundgesetz, also forbids segregation of pupils according to the means of their parents (the so-called Sonderungsverbot). Therefore, most Ersatzschulen have very low tuition fees and/or offer scholarships, compared to most other Western European countries. However, it is not possible to finance these schools with such low tuition fees, which is why all German Ersatzschulen are additionally financed with public funds. The percentages of public money could reach 100% of the personnel expenditures. Nevertheless, Private Schools became insolvent in the past in Germany.'

Speaking generally, the only variety of private schools for elementary to high school in this region with a hint of acceptability are those of the Waldorf variety (German text - ). And even then, it is generally connected to the idea that anyone in a private school is not capable of handling a normal education as provided by the public system (though in the case of the Waldorf schools, this is seen, at least potentially, as an advantage, as not everyone fits into a single system).

And only now do I read the headline - obviously, this information has little to do with a developing country.

Though whether this sort of information is meant to influence the policy debate concerning public/private education in the U.S., and whether the U.S. can be considered a developing nation, is left up to the loyal and disloyal readers to decide.

I'm happy to see BRAC getting good press - they have been an important organization in Bangladesh.

In my observation, the market it is a bit over-saturated with private schools in Bangladesh. The culture seems to have bought into the (greatly over-blown, IMHO) idea that education is the way a country can climb out of poverty. Sure, education is important. So is technological innovation. So is the absence of corruption. So is political stability. So is capital savings and retention. Etc. etc.

Lots of things are important. Sometimes it seems as though a lot of people believe education is everything, from which everything else follows. Dhaka is full of well-run and well-attended schools. If all it took were education, Bangladesh would be Singapore by now. Obviously there is more to it, which is something that makes BRAC so important; they educate, but they also do many other things.

More than 80% of Haiti students go to private schools. From my own observations, the best schools there tend to be non profit, private and run by catholic religious orders. The public schools are somewhere in the middle in terms of outcomes. The schools with the worst records tend to be small, private and for profit. Tyler had a post that touched on this issue 2 years ago:

Rosenberg sets up a two part article praising two very different kinds of private schools in developing countries. In part one she discusses BRAC, which is Montessori and dislikes tests and does lots of singing and dancing and believes in everything that's hippie and holistic. And it gets great results. Rosenberg says she'll discuss the successes of the other approach, being carried forward by Peterson, in part two.

The top two NYT comments both bash testing and Peterson and how they corrupt everything noble about learning and crush puppies. Predictable really.

Private schools may perform better than public ones, but is this relevant?

Yes, this is a crazy thought, but at least in India - I've been through the schooling system there - the end-of-school exams are highly predictable, and although I officially went to a private school, I also had a private tutor. I learned enough to know that if I practice on the last 5 years papers, I would be able to solve 90% of the questions that would eventually came.

I didn't learn any of the subjects, I simply learned what questions to expect.

And I am part of the elitest elite. My guess is this is what private schools do - they drill you through all possible questions that can come in the exam. Which is fine given the way everything is at the moment, but I would hesitate greatly to say the results demonstrate private schooling is more efficient than public. It does so, given a very predictable set of exam questions.

Of course, there is no question that doing well on an exam enables better outcomes - better colleges, for instance - but is this what school is for? Aren't students supposed to think and challenge themselves, instead of learning how to game the system?

Public schools do the same. So that's hardly a distinguishing factor.

In any case, you still ended up doing fine, right? In spite of the drilling, predictable etc. that you loathe. Sometimes the drills etc. get a far worse reputation than they deserve. Even if teaching to the test won't make stellar thinkers and scientists it is perfectly adequate for the majority that ends up doing prosaic jobs any ways. In any case, far better than the worse option of functional illiteracy.

I'd love a quixotic system that taught everyone to " think and challenge themselves" but looking at the baseline that's hardly our biggest problem right now.

In Delhi, public schools pretty much don't do anything.

Rahul, you might want to find out more about Singapore's 'A' Level Examinations. We still have a fair share of 'drill-and-kill' questions, but we are moving towards 'higher-order-thinking' (HOT) questions that are by their very nature impossible to prepare for. Our chemistry paper is the best example of this move towards HOT questioning.

The students just have to absorb the new information, process it, discern the patterns, synthesize it with their prior knowledge, and finally apply the pattern to a given question. On the spot.

It truly separates the great from the merely good. Though we are trying to game the system by exposing students to more questions of this sort to 'force' open their neural pathways. Not sure how it's working out so far though.

Campbell's law in action.

Think about what you're saying: You paid a private school to advance you further into the world of (eventual) job acquisition, and the private school delivered on its promise.

Would you rather have gone to a public school (or any other school for that matter), learned to think and challenge yourself, and then walk away with a great mind at the expense of your status as part of the "elitest elite?" I know we are all supposed to thirst for knowledge for its own sake, but compared to abject poverty, that is a bit of a luxury.

I agree, as I don't think that saying private schools can "perform" better than public schools is relevant at all. It all depends on the student himself/herself. Even though that it's true that private schools kind of force more on you academically, it doesn't necessarily mean that the student will do better. As with private tutors, additional classes, etc., it doesn't necessarily guarantee a better "outcome" as the blog stated. Nevertheless, it doesn't matter if private schools are instituted into the poorer countries because at the end of the day, it comes down to the availability of academics for the students, which in this case, they already have, and their own private willingness to learn and succeed.

There's a weird thing talking about causal relationships between teacher wages and student outcomes. On the one hand, if you have private school teachers being paid less than public school teachers in India, and the private school teachers get better outcomes, that might signal to me that more teachers want to teach at the private schools because they expect more disciplined and more interested children. Public schools may pay more because they required more experienced teachers that can handle a classroom on fewer resources or with poorer children.

Instead of a causal relationship where wages leads to results, you may actually have a relationship where results impacts wages, because the results change the supply/demand relationship for teachers that want to come to the school. On the other hand, if every school in India - both private and public - suddenly increased their wages by 20%, how many would-be doctors and scientists and accountants would go into teaching instead of other degrees? That could potentially bring more talent to the educational system and increase results. It's probably not an either/or scenario.

I'm very happy to discover this web site. If you don't mind Could I also reveal a tip. Fidning a sexy muscles nude man could well be dificult, in the event you reside in Sourthern California and you also have to have a sexy muscles love firm, mouse click my link.

Comments for this post are closed