Regarding the first point, the conclusion is based on a selective reading of the research. For example, OECD’s PISA report, which does not find a correlation between competition and achievement, is cited frequently. What does that prove? Not much. The PISA report enquires into cross-country correlations, but it cannot tell us anything about causality. It is thus conspicuous that Muir ignores all cross-national academic research on PISA/TIMSS scores – in which one of the most consistent findings is that private-school competition improves international test scores. The best strategy finds that a one standard deviation (SD) increase in private-school competition raises mathematics PISA scores by 0.3 SD while decreasing costs by 0.4 SD. Not too shabby, since most countries do not even have bona fide choice programmes in place. One has to ask why Muir ignores these findings.
Another problem is Muir’s reading of the Swedish research, which he argues shows that there are no long-term positive effects of competition. He might be interested in the most recent research by Böhlmark and Lindahl, which is omitted. They find positive long-term effects of competition in compulsory education. I have previously provided a longer discussion of this paper.
We can also see evidence of selective citing from the British evidence. For example, a well-cited paper by Gibson, Machin and Silva found the overall impact of competition to be minimal – while showing a moderately positive effect among voluntary aided schools, which have some autonomy to act upon competitive incentives. In another paper, Silva cites this finding as some evidence ‘that increased parental choice and school competition, coupled with some degree of school “autonomy”, can lead to an improvement in standards in education’ (p. 19). Muir mentions only the overall null effect and omits the positive finding among voluntary aided schools.
This failure to note nuances in the research demonstrates a crucial point: choice opponents rarely understand the choice programmes they cite. In general, a majority of studies worldwide find moderately positive effects of competition. Muir argues that the impact is small and sometimes not significant. But what should we expect? Almost all programmes are small, target a tiny proportion of pupils and come with heavy restrictions and poor system design. One cannot just point to research that purportedly is evaluating competitive systems. One has to analyse the systems to understand whether they increase competitive incentives and how the incentive structure is shaped. Such analyses can help explain why the Chilean evidence is mixed while the Swedish research is consistently positive.
What about the second point? Muir tells us that for-profit schools often perform worse than non-profit schools, partly by referring to my IEA study. In my case, it is simply not true. In the instrumental variable model, which Muir ignores, the differences between for-profit and non-profit free schools are erased entirely.
In hindsight, I am sceptical about what my paper says about differences between municipal and for-profit schools because of selection bias. But since this probably favours non-profit schools, and since my instruments are more likely to be valid for comparisons of different types of free schools, the conclusion of tiny quality differences between these types remains appropriate.
Muir also ignores research from Chile which uses better methods than the only study he cites. Elacqua finds that for-profit chain schools, but not other for-profit schools, perform on a par with the best Catholic schools, while performing better than some other non-profit school types. A key argument for profit-making schools is that they often scale up in chains; it is peculiar that Muir has omitted this research.
Most importantly, the report overlooks the importance of a supply-side dynamic for competition to function. Schools need to start up close to failing schools so that pupils have realistic alternatives. As Silva argues, ‘policies aimed at increasing school choice [at least in the current English setting] have been unable to sever the link between residential location and school attendance’ (p. 31). The profit motive gives strong incentives for actors to enter the market and replicate success, which is found in Chilean research. It is partly because England already has a non-profit sector that it is unlikely that non-profit schools will provide the supply necessary to free up choice significantly.
Regarding the third argument, while we are provided quotes from social theorists who don’t like profits, we are not presented with any evidence. It is impossible to provide an argument against it, simply because no real case is made for it. Muir argues that the case for profit-making schools is ideological, but, in fact, his final argument is almost solely an ideological case.
There is one argument in the final section, however, which deserves discussion. Muir contends that profit-making institutions might be incentivised to improve test scores, but that education is about more than test scores. It is extremely difficult, Muir argues, to contract companies to provide complex quality that cannot be measured by test scores.
Yet, if it is not measured, measure it: we need a robust information system to complement choice. Apart from league table reform, school inspection scores and parental satisfaction surveys should also play a role. Research (here and here) finds that Ofsted inspections produce better outcomes. It also shows that inspection scores explain parental and pupil satisfaction levels over and above school test scores. These, and other, gauges would incentivise profit-making schools to focus on non-test score quality.
Apart from this, the report has little to offer those of us who are not already convinced that profits are bad for education. All proponents of choice should take good arguments against profits seriously. The IPPR report, however, fails to provide a persuasive case why for-profit free schools should be banned.