Private Schooling In India: Results from a Randomized Trial

Karthik Muralidharan runs very large, randomized controlled trials on education in India. His previous work showed that performance pay for teachers in India has large and significant improvements on student learning. In his latest paper (with Venkatesh Sundararaman) he reports on the results of The Andhra Pradesh School Choice Project, a long-term randomized controlled trial covering over 6,000 students in 180 villages for four years (2008-2012). The study offered students a lottery for a private school scholarships and lottery winners were compared with non-winners. The results show modest improvements in learning for private school students and big increases in school productivity.

We find that private school teachers have lower levels of formal education and training than public-school teachers, and
are paid much lower salaries. On the other hand, private schools have a longer school day, a longer
school year, smaller class sizes, lower teacher absence, higher teaching activity, and better school
hygiene. After two and four years of the program, we find no difference between the test scores of
lottery winners and losers on math and Telugu (native language). However, private schools spend
significantly less instructional time on these subjects, and use the extra time to teach more English,
Science, Social Studies, and Hindi. Averaged across all subjects, lottery winners score 0.13 σhigher,
and students who attend private schools score 0.23 σhigher. We find no evidence of spillovers on
public-school students who do not apply for the voucher, or on students who start out in private schools
to begin with, suggesting that the program had no adverse effects on these groups. Finally, the mean
cost per student in the private schools in our sample is less than a third of the cost in public schools.
Our results suggest that private schools in this setting deliver (slightly) better test score gains than
their public counterparts, and do so at substantially lower costs per student.

As Karthik notes in a Ideas for India short article that summarizes:

Since private schools achieved equal or better outcomes at one-third the cost, the fundamental question that needs to be asked is “How much better could private management do if they had three times their current level of per-child spending?”

Is any economist doing more important work with greater potential for real improvement in the lives of millions than Karthik Muralidharan?


'The results show modest improvements in learning for private school students and big increases in school productivity.'

So, how does one measure 'schiool productivity' if is not based on something other than student learning?

Or is the sort of question that a tenured professor, who can ignore any and all student and peer evaluations using any standard of measurement, simply does not consider?

Easy answer to your question PA: " better school hygiene". There is a hypothesis, that is backed by some data (see here: ), that better hygiene and nutrition will increase IQ. So those private schools that are washing student hands and feeding them better (since the school day is longer, and they don't go to their worm-infested shacks like the public school kids) are making their pupils smarter.

NB.- note how dumb the Third World types are--a fact you would readily see even if you don't agree with Steve Sailor as I don't, if you spend any time overseas like I do. Borderline retardation is 80 IQ and note most of Africa is in the 60s and where I'm at now is only 86 (Philippines). Patience is needed when dealing with the natives, as they really are stupid not just trying to be cool as is sometimes the case.

One might argue that the gains in this case are attributable to the students not having to fight off as big a microbial load. Of course, that's boring crap like cleaning supplies, janitors, functioning toilets, not interesting stuff like complicated matrices of teacher/student ratios, study hours, etc.

Which leads to my perennial question, why don't all the foreign aid-busybodies wandering all over the Third World concentrate on sewer systems and potable water first, then we can fret over whether they have iPads.

A hypothesis of mine is that the shortness of a population is indicative of its IQ increase potential due to nutrition, since there is relatively little genetic variation across populations when it comes to height, ex-exceptions like Pygmies and Negritos.

So the Chinese, South East Asians, Indians, and some African countries would have ample room for IQ growth.

What sticks out to me when traveling in Latin America and SEA is not so much how dumb the people are, but the lack of smart people (e.g. IQ 130+). After all, we have plenty of dumb people in America as well, so I'm used to it.

@Kabal - Agreed. And I have nothing against stupid people, since we all are stupid to various degrees, but here in the Philippines there are some really stupid ones. Like the tout that tonight ran behind me, for about 10 minutes, hollering "hey Joe", "hey hey hey", "hey Joe", trying to get my attention to buy something. Could not take a hint until they ran out of energy. Like the scene of the chasing pirates in Captain Phillips in the first failed attempt to board the tanker. Taxi drivers will do the same thing: they will stop and honk at you, imploring you to stop walking and get in their cab, when you are walking in the street, and after you waive them off they will still slowly stalk you in case you change your mind. Stupid, stupid taxis.

I once saw an article that said that a average professional in the USA would be a superstar in a developing country, while a superstar in a developing country would only be average in the USA. Sounds about right. Again, I love this place which is in many ways better than the USA and probably will marry a local and stay here, so don't accuse me of racism, just making some observations.

"So, how does one measure ‘schiool productivity’ if is not based on something other than student learning?"

To measure productivity you need to know the inputs and the outputs. In the case of schools the output is learning and the input is money. Example: School A and School B have the same student learning outcomes but school A spends half as much money per student as School B. As School A achieves the same output with fewer inputs we can say that School A has higher productivity than School B.

School productivity is measured as learning/$. In this case, the private schools achieve somewhat better learning outcomes at 1/3rd the cost thus over 3 times the productivity.

Yes, looked at that number, it should be analyzed really carefully. First, Mr. Muralidharan research was not about the costs of public and private education. He cites Tooley, James (2009),The Beautiful Tree: A Personal Journey into How the World’s Poorest People are Educating Themselves (New Delhi: Penguin) as the cost data source. The mean cost per student, more or less, is the total annual expenditure in education institutions divided by the total number of students. Based on this simple definition private education has some cost advantages over public one totally unrelated to the education process: a) private schools don't have the fixed cost of retired teachers while public education certainly does, b) public education pays for educating teachers in first place, c) public education has a big bureaucracy in charge of lesson planning, student grades processing and storage, etc. What would happen if private schools paid for retired teachers, the formation of teachers and some other costs?

Indeed. Further, he seems to have used the public school spending number (~Rupees. 8390 per student )from this study by Ambrish Dongre titled "What is the Per Child Expenditure in Government Schools?"

Unfortunately that's an average number across the whole state of Andhra Pradesh (rural + urban) whereas this study focuses on rural schools alone. Mr. Muralidharan seems to be obviously using an inappropriately inflated estimate of public school spending.

Another big flaw is that he's using self-reported income numbers from individual private schools in this study. Not only does that skew the rural-vs-urban aspect (i.e. comparing a state average public school metric versus a rural only private school metric) even worse it ignores the practical reality that almost *every* private school in India needs a significant, un-receipted, cash-only "donation" for tax evasion purposes. There's an obvious conflict of interest that induces a private school owner to under-report his income.

In effect he's underestimating private school spend & over-estimating public school spend. Now I hate to presume Karthik Muralidharan is making such obvious errors, so for now I'll assume I'm misreading something. Would love to be corrected.

There seems to be a really good result backed by data : "We find positive effects of vouchers on test scores on all of these subjects (large and significant for Hindi). Thus, adjusting for instructional time, we see that private schools are more productive because they are able to deliver equivalent outcomes as government schools on Telugu/ Maths even with substantially less instructional time, and used the extra time to deliver better outcomes on other subjects (especially Hindi)." Problem is that he got carried away when using other people's results on costs.

“How much better could private management do if they had three times their current level of per-child spending?”

Ths supply of good teachers is about fixed, this'll just increase the cost of schools, with near zero benefits in quality.

I won't deny there is a law of diminishing returns, but I suspect your point is more important in the West than India. Note that the private schools had lower pay and longer hours for teachers. At a guess, this suggests more teachers could be found if some of the extra money went on is pay was increases.

Who is this Ray Lopez character and why is he so deeply unlikable?

Apparently 24 year old Filipinas like him a lot. And they're super-hot, just ask him.

I love how they run an interesting, complicated trial, and then someone can't help jumping straight to the, "now just imagine if private management had more money!" Fail.

Work like this is wasted when we treat markets as magical black boxes, instead of trying to understand exactly how their incentive structures work (or don't work).

Indeed. I think it's somewhat funny to ask "How much better could private management do if they had three times their current level of per-child spending?"

Obviously, we are not talking about private schools suddenly increasing their fees 3x overnight. So implicitly this is asking for Govt. funds poured into private schools? Problem is, once it's not pinching the parent's pockets a lot of the efficiency will vaporize. Parents hardly care if the school wastes money that's not coming from their pockets.

Let's start with the assumption that the answer is NOT "they'd be 3 times as good," which is what most people's minds would naturally jump to.

I am sure that caste has everything to do with the results.

Please study what a randomized controlled trial is.

One worrying aspect about this study is the potential conflict of interest. It sounds like much of the actual groundwork (enumerators, surveys, trackers, assigning vouchers etc.) was done by employees of the Azim Premji Foundation (one of India’s leading non-profits working on education).

Now I can see plenty of reasons why the converse conclusion (i.e. Govt. schools better) would hurt the Azim Premji Foundation's interests.

While I would obviously put great trust in a study by the education bureaucracy that reached the conclusion that private schools were better and cheaper, why would I expect that study to exist (or be publicized if it existed)? Usually the people doing research are going to have some interest in what they find (if nothing else, more grant money if they do "interesting" research).

It's like I'd trust Alex Tabarrok to tell me whether FDA regulation is too onerous or not but probably not a study run by Pfizer nor an FDA scientist.


You may want to read these two pieces by Anurag Behar, the CEO of Azim Premji Foundation - "Myths of Privatopia" ( and "The ideology of education" - ( - to get a sense of the Azim Premji Foundation's views on govt. schools versus private schools and then decide if there's any conflict of interest at play.

So, the study seems to be internally valid as the lottery applicants were randomized. Is it externally valid? Could there be bias in who applied? Or, is that the issue addressed by "no spillover"? - Genuinely curious, thx

And by internally, I mean with respect to Indian students in the Indian school system, to be accurate.

The real question is “How much worse could public management do if they had one-third their current level of per-child spending?”


again, i can guess the study results from the title of the post.

But I bet you can't guess the same when Prof. Cowen writes a post title - 'John Roemer changes his mind on a bunch of things, including socialism,' for example, is quite unlikely to lead one to the conclusion of John Roemer himself - 'The social-democratic model remains the best one, to date, for producing a relatively egalitarian outcome, and it relies on solidarity, redistribution, and private ownership of firms.'

Has there ever been any doubt regarding that specific conclusion?

Was Soviet Russia more or less egalitarian than post-Soviet Russia?

Which system was better?

To say that social democracy is more egalitarian than capitalism is practically tautological, given that the defining difference between the two systems is the size of their handouts.

Not sure how India runs their public schools, but in the US I'd guess that special ed programs and athletics increase the public ed bottom line. Maybe a more fair question might be, "what could private schools do with three times their funding - as long as they provided the equivalent opportunities to all comers that public ed does?"

No, my understanding is that unlike in the US, in India public schools do almost little or nothing with regard to Special Ed and athletics. Typically only expensive private schools ( not those included in this study) do much for special Ed and Athletics.

This supports the model of PURELY private schools that I advocate:

Comments for this post are closed