The conclusive case for school choice
On 4 October, Karthik Muralidharan of the University of California, San Diego and World Bank’s Venkatesh Sundararaman released a paper, titled The aggregate effect of school choice: evidence from a two-stage experiment in India, dropping a gigantic bombshell on the cobwebs engulfing education policy in India. Their paper was based on the Andhra Pradesh School Choice Project that was based on randomized control trials (RCTs), for which Massachusetts Institute of Technology’s Abdul Latif Jameel Poverty Action Lab (J-PAL) has come to be well-known.
Before discussing their results, let us understand what RCTs are, and what school choice is. An RCT tries to have reasonably large control and sample groups, with its constituents randomly chosen and differing in only one way: the treatment, or the intervention. This way, the law of large numbers suggests that any significant difference between an output parameter post-intervention for both the groups should be attributable to the intervention. In this way, the theory goes, one can avoid complicated statistical analyses using randomized controls and avoid caveat like ‘correlations are not causations’ which haunt other kinds of economic studies. There are criticisms of RCT, the most significant one being the high cost of running of such trials, because of which experiments are often localized, thus yielding results that cannot always be extrapolated nationally or internationally. Having said that, RCTs are a valuable research tool in the field of development economics.
Vouchers allow poor students options beyond the free government school by compensating them if they want to attend a private school instead. School choice is a broader concept. It also includes charter or similar government school models where teachers could be fired more easily. It could also include the concept of government-aided private schools, where the aid is not necessarily on a per-student basis. The broad idea is that having more market-like structures in education leads to competition between schools, greater choice for students and parents, and incentives for teachers to show up and perform. (For more on school choice in the context of India’s free and compulsory education law, see our earlier piece.)
Nobel laureate economist Milton Friedman had come up with the idea of school vouchers. James Tooley’s book The Beautiful Tree popularized the existence of low-cost private schools. Often students are not allowed to top-up these vouchers, which would allow them to go to more expensive private schools, and schools are required to accept students on a first-come, first-serve basis (so they cannot skim off the financial or intellectual cream). In Muralidharan and Sundararaman’s study, these conditions were applied.
The first thing they found is that those students who attended private schools by using the voucher they won in a lottery performed marginally better. There was “no difference between the test scores of lottery winners and losers on Math and Telugu. 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”. On an average, across all subjects, lottery winners did better. But a statistically significant yet small improvement may not be much to write home about as far as the case for school choice goes.
The significant result is that the average cost per student in private schools was just a third of the cost in public schools, suggesting 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”. To emphasize that, suppose a rural government school spends around Rs.1,500 per student per month in Andhra Pradesh (a reasonable estimate). Then, an average private school in that area charging students Rs.500 per student per month is actually delivering slightly better results on average across subjects. How do the private schools manage this seemingly incongruous result?
In one word, incentives. Among other factors, private schools pay their teachers lower salaries, while making those teachers work harder. This is because jobs in private schools, unlike government schools, are not secure. If teachers do not perform, they will be fired.
Now consider if the government were to give all (or only the poor) students a Rs.1,500 monthly voucher which they could redeem at a government or a private school. The output of academic teaching at private schools—which is already slightly higher to begin with—will only go up because they will now have more resources.
Educational outcomes and quality will not go up linearly with increasing resources, of course. The amount of the proposed voucher programme could be in the middle of the average cost of attending a private school and that of attending a public school—and such a policy reform improves outcomes over the present arrangement, while saving substantial taxpayer’s money. Rupee for rupee, private schools outperform government schools by a large margin. And even if we were to raise government funding and the size of the voucher, school choice would still make sense.
The “how much” of when it comes to educational and other subsidies is a political and philosophical debate, whereas the “how to” is simply a matter of being smart and efficient. Indian politicians, especially of the leftist persuasion, routinely try to fool voters and analysts by bragging about how big budget allocations and expenditures are, instead of focusing on what results a particular welfare scheme has delivered. It is outcomes that matter, not outlays.
The usual caveats of RCTs apply—what is true in Andhra Pradesh may not necessarily apply to the rest of India and so on—although of course, the sheer size of the cost-adjusted outcomes is extremely noteworthy. So to be sure, let us evaluate some more evidence. In an April 2013 paper for the Friedman Foundation for Educational Choice, Greg Forster found that of the 12 American RCTs about school choice, 11 showed improvements—six that benefit all, five that benefit some students, one with no impact, and none with a negative impact.
Twenty-three empirical studies employing all methods (not just RCTs) examined school choice’s impact on academic outcomes in government schools—22 found it improves public schools and one finds zero impact. Six studies examined school choice’s impact on taxpayers—all found that it saves money and none found a negative impact. Then one can peruse the monograph, titled Educational vouchers: Global experience and India’s promise, prepared by New Delhi-based think tank Centre for Civil Society, which compiled evidence from countries as diverse as the US, Chile, New Zealand, Hungary and Denmark.
The evidence for school choice (i.e. adjusted for money spent, educational outcomes are better at private schools than government schools) is overwhelming and robust. But government teacher union leaders represent a powerful interest group that blocks change, and they are supported by the outdated mindset that education should not be a profit-seeking industry.
There is only one real reason why profit continues to be illegal in parts of the education industry—because profit in modern India continues to be a dirty word for some Nehruvian ideologues. Never mind that these ideologues turn a blind eye to politically-connected incumbents who want the education sector to remain off limits to profit-making. It’s the classic case of bootleggers and baptists coming together to ban alcohol.
Private schools have longer school days, fewer holidays, lower teacher absence, lower pupil to teacher ratios and less cross-grade teaching. They provide better school facilities (more toilets, for example) and yet spend far less. Yes, the teachers at private schools are less-educated and less-experienced, but evidence from around the world shows that incentives to perform are much more important than mere credentials.
That school vouchers are yet to gain unqualified endorsement from India’s intellectual establishment is extremely unfortunate and detrimental to the future of millions of children—the meekness of policy analysts, inexplicable as it is, only gives cover to the self-serving bootlegger-baptist coalition in the education sector.
Moreover, centralized and statist systems are hardly innovative. As mobile computing and the Internet transform our lives, our education system should be opened up to create more incentives for technology-based competition too. Thousands of government schools run empty and they continue to receive subsidy. Have we forgotten the price-decreasing and quality-increasing impact of such competition in other industries?
India needs to break the stranglehold of the state on schools, hold teachers accountable and dissolve the dictatorship of the salariat, to use economist Pranab Bardhan’s memorable phrase.
Rajeev Mantri and Harsh Gupta are co-founders of the India Enterprise Council.