Syndicate content

Private Schools or Public? Justin Sandefur responds to Kevin Watkins

Duncan Green's picture

Everyone enjoyed last week’s arm-wrestle on public v private education, so in a titanic struggle for the last word, Justin Sandefur (right, in the private corner) and Kevin Watkins (in the public one) are back for another go. Seconds out, round two…..

Dear Kevin,

Thanks for your reply. You are of course quite right that I wear a Pearson corporation logo on a chain around my neck to ward off evil spirits, I can’t stand (or understand) solutions with multiple steps and regularly visit my local medium to have a chat with the sadly departed Milton Friedman.  But despite all that, I want to contend that we agree on almost all the necessary ingredients for a constructive policy discussion.  I’ll end with where I think our core disagreements are.

Evidence over ideology

First, we agree the debate should be based on empirical evidence, not ideology.   In an essay entitled “Policy Analysis with Incredible Certitude“, economist Charles Manski re-examined Milton Friedman’s dubious advocacy for private-school vouchers in the U.S., noting:

“Friedman cited no empirical evidence relating school finance to educational outcomes.  He posed a purely theoretical classical economic argument for vouchers…”

“Rhetorically, Friedman placed the burden of proof on free public schooling, effectively asserting that vouchers are the preferred policy in the absence of evidence to the contrary.  This is the rhetoric of advocacy, not science.  An advocate for public schooling could just as well reverse the burden of proof, arguing that the existing educational system should be retained in the absence of evidence.” 

What Manski calls a purely theoretical argument was, in effect, just an expression of faith and political ideology.   I think we both agree that the wisdom of investing in voucher programs or similar schemes should hinge on evidence and facts.

Shared values

Second, I’m going to venture we agree on the underlying goals or values that should guide our evaluation of the evidence. I’d list three.
#1.Access.  Primary schooling should be free, always and everywhere. 
#2.Quality.  Governments have a responsibility to ensure quality education for all children. 
#3.Equity.  Equality matters in and of itself, beyond mere improvements in, say, average learning outcomes.

A good expression of these values came from Pauline Rose via twitter.

If there’s any disagreement between us on these three values, I suspect it’s on the definition of “ensure” in point #2.  I’m happy for it to mean “pay for”.  But I noticed that when the UNESCO Education for All report discusses public-private ventures, it states in bold, colorful text: 

“The bottom-line obligation for all governments is to develop publicly financed and operated primary schooling of good quality for all children.”

This a priori insistence that governments not only finance, but directly operate primary schools essentially cuts off any evidence-based policy debate on public-private partnerships before it can begin.  (Friedman was never quite this bold!)  It also suggests that we artificially limit the tools at our disposal to pursue our core goals of access, quality, and equity.

Growing empirical consensus

Third, we agree on most, though clearly not all, of the relevant empirical facts.  Let’s see if we can hone in on where we disagree by revisiting the three issues I started with in my first post. 

The crisis. I think we’re in full agreement that the current state of education in many developing countries is unacceptable.  I focused on low learning levels, but you’re right that there’s a long way to go on enrollment in many countries, not least Pakistan.

Affordability. You noted that it costs about 1/10th the minimum wage to pay for a low-fee private school in Lagos.  In the blog comments, Ruth Nyambura did similar calculations for Kenya, showing that even low-cost schools are very expensive for typical Kenyan households.  This is all sadly true, yet potentially irrelevant given we agree that primary access should be free. 

Parents should not be paying these fees.  If a country decides to pursue any form of public-private partnership for primary schooling, we agree that the state should, at a bare minimum, foot the bill for school fees.   The relevant point here is that fee levels in low-cost private schools are much less than the cost (to the state) of operating public schools.   From a public finance point of view, private schools cost negative money if the alternative is state provision.

Public vs. private performance. This is where we disagree on a few matters of fact:

  • You insist there’s no evidence that charter schools in the U.S. are working, despite multiple randomized trials showing significant gains.  These trials show charter schools are most effective for non-white pupils in disadvantaged urban areas who are low-achievers at baseline, but less effective for suburbanites.  Instead you cite a Stanford report (CREDO) that has come under considerable methodological attack for its reliance on observational data and potential methodological flaws.  Even so, CREDO  notes that “For students that are low income, charter schools had a larger and more positive effect than for similar students in traditional public schools.”  CREDO also argues against attempts to cap charter school growth.
  • My fault for introducing Sweden to the discussion, which is probably of limited relevance to the countries we’re debating, but the blog post you linked to on the Swedish model states unequivocally, “There is no question that, at a system-wide level, the presence of free schools improves outcomes.”  The equity effects here are more complicated, though your link cites research showing, “students from low-income families benefit more than those from high-income families”.   I’m not sure either of us wants to rest our case on Sweden.
  • More relevant is the Kenya example, where you worried that the results in my co-authored work showing higher performance of private schools were driven by elite academies.  There’s ample evidence this isn’t true.   Two-thirds of private schools operate on lower budgets than the median public school, while roughly 85% of private schools score higher than the median public school. 
  • Finally, you shouldn’t have to take it from me.  Let’s turn to South Asia, where the UNESCO EFA report (p. 166) sums up the empirical debate:


“There is evidence that in many contexts private schools are outperforming state schools. In parts of India and Pakistan, children enrolled in low-fee private schools perform better, on average, than those in government schools, once adjustments are made for socio-economic status and other variables (Andrabi et al., 2008; Aslam, 2007; Das et al., 2006; Muralidharan and Kremer, 2006; Schagen and Shamsen, 2007).” 

There is increasingly little wiggle room for debate on this.  So, where do we go from here?

Fixing failed states

Picture a village in rural Kenya with two schools: one government school with multiple classrooms but chronically absent teachers and abysmal academic performance; and one private school with motivated staff, involved parents, and an accountable management structure producing higher scores.  But the private school charges fees that impose a serious economic burden on poor families.

What’s the right response here?  Whoever runs the Labour Campaign for International Development twitter feed picked one of the more popular lines from your post:

I can’t help wondering who this imperative is directed at.  If the “fix the state” message is directed at the parents in my hypothetical village, it seems like a cruel, Marie-Antoinette-esque joke.  And if it’s directed at DFID or any foreign donor, well, Matt’s response seems appropriate.

The proper audience — though I doubt many are listening to us — may be national governments and Ministries.  I’m just not convinced we have clear solutions to offer them.  My co-authors and I spent years working with the Ministry of Education in Kenya trying to identify what works, and we’ve pretty much failed so far.

I know you’re more optimistic. You mentioned a McKinsey report (co-authored by none other than Pearson’s Michael Barber — he’s everywhere!) as providing powerful examples of how investing in state systems can improve learning.  But it’s not clear this report is relevant to rural Pakistan, India, or Kenya.  The full sample of  McKinsey’s “sustained improvers” consists of England, Hong Kong, Korea, Latvia, Lithuania, Ontario, Poland, Saxony, Singapore, Slovenia, Boston, and Long Beach (USA).  

I’m left sympathizing with Claire Melamed from ODI.

What if we can’t fix failed state systems anytime soon, as most evidence would seem to suggest.  What do we do in the interim?

First, we should be trying as hard as possible to come up with solutions that provide affordable, equitable quality education across the world –and that is going to take a lot of experimentation and monitoring. 

But, second, we should avoid the urge to squash the private sector.  This urge is reflected in Kenya’s decision to penalize private school graduates in secondary school admissions, and in some of the more controversial pieces of India’s Right to Education Act.  Private schools are laboratories for approaches that can be applied in the public sector, and can be an important tool in extending access to quality education. 

No one is calling for mass privatizations, whatever that means.  I don’t think we currently have a sufficient evidence base on the impacts of vouchers in low-income settings to even justify rapid expansion of these programs.  But the UN Declaration says everyone has a right to education.  Sadly, in many parts of the world today, send your kids to a government school, and they won’t get educated. We shouldn’t be taking away kids’ human rights based on an ideological opposition to the private sector.

I’m sure we don’t agree on everything here.  But I do hope we have real common ground about the terms of an evidence-based policy debate, the values we’re pursuing, and the kind of research that remains to be done.


Kevin’s already sharpening his pencil...


This post was originally published on From Poverty to Power

On Monday, August, 13, we will publish Kevin Watkins' response.

Follow PublicSphereWB on Twitter


Submitted by Mark on
In as much as I believe that private(capitalist) systems tend to allocate resources better and manage the said resources in an optimal manner, the case of a school being firstly a social institution as opposed to a commercial one cannot be overstated. Privatizing every underperforming institution is throwing a less capable model at a solid model that just happened to trip on itself. The detachment from the teachers towards the students and the parents stems only from the fact that the employers of the teachers are an amorphous body called the government that is centralized. In the case of Kenya this is bound to change with introduction of the counties. More local more accountable. Private is not necessarily better.

Add new comment