Six founders of NCEE’s Superintendents Alliance make the case for proficiency based education. Read more.

By: Marc Tucker

TheMythofMarketsinSchoolEducationBen Jensen, education director of the Grattan Institute in Australia, recently produced the best analysis of the way markets do or don’t work to produce improvements in school performance I have yet seen.  Overcoming the methodological difficulties that have plagued such studies in the past, he makes a well-documented case not only that market strategies have failed to improve the performance of schools at the system level, but that, given the nature of this particular market, they cannot do so.  He describes why this market fails to produce the market forces that would produce incentives for schools to improve their performance and he then explains why it is so hard for democratic governments to make policies that would overcome those sources of market failure.

This is, of course, no small matter in the United States.  The expansion of the population of charter schools is right up at the top of the agendas of both the Republican and Democratic parties in this country.  While many people believe that, in a democracy, parents ought to be able to choose their child’s school irrespective of the impact of charters on student achievement, charters have most often been justified on what is taken to be the self-evident proposition that competition in the school market will produce better schools at lower cost, as competition has done in so many other markets.  But what if that is not true?  What if this particular market does not function that way?  What if there are so many sources of market failure and those sources of failure are so hard for government to fix that increasing competition among schools not only will not but cannot lead to improved school performance overall?

Jensen begins his report with a thorough review of the literature on this topic, including the studies done by American researchers.  While he notes that the effect sizes claimed by researchers on both sides of the controversy are very small, he also notes significant differences in findings, which he attributes to methodological difficulties, which he then describes in some detail.

Taken together, these methodological problems make it very difficult to determine how much competition there actually is in any given situation, much less the degree to which parents make their choices on the basis of relative school performance and even less the degree to which competition on school performance leads to a net improvement in student academic performance overall.

So Jensen and his team decided to set up a research project in the greater Brisbane area in Queensland, a state on the east coast of Australia, a project designed to enable them to overcome all of the methodological problems he describes with previous studies.  Brisbane is one of the largest metro areas in Australia.

It turns out that Australia is a good place to do such a study.  According to the OECD, 36 percent of Australia’s schools are private, compared to an OECD average of 14 percent; government funds 78 percent of the costs of these schools, compared to an OECD country average of 57 percent; 35 percent of Australian parents say that they chose one of these private schools for reasons other then the fact that is was the closest school, compared to an OECD average of 25 percent.  And, finally, more Australian principals report that they feel competitive pressures from other schools than in any other OECD country.

CThe question the study addresses is the degree to which competition among schools can lift the performance of schools across the system.  Answering that question requires much more nuanced measures of the competitive pressure experienced by schools within the system than previous studies have used.  Jensen and his colleagues looked at the structure of local school markets, school performance, affordability, capacity and enrollment patterns.  The model they built enabled them to vary such factors as the distance of possible competitors from any given school, how many schools were within a given distance whose performance exceeded that of the target school, how much the performance of the school was exceeded by any given potential competitor school, the amount of school fees charged by competitors, how close possible competitors were to their full enrollment capacity, and so on.  Using this dynamic model and the data they were able to obtain to populate that model, Jensen and his colleagues were able to calculate the competitive pressure on a given school from other nearby schools.

If the thesis that competitive pressures lead to improved system performance holds water, one would expect to find that the more competition a school experiences, the better it will perform.  One would expect to see parents moving their children from low-performing schools to higher performing schools.  In time, one would expect the lowest performing schools to disappear and the next-lowest-performing schools to experience serious enrollment declined.  Conversely, the better schools would be expanding in both size and number.  So goes the theory.

But that whole logic line assumes that there is enough competition among schools to drive the choices that lead ultimately to higher school performance.  But that turns out not to be the case.  Under very conservative assumptions (favoring the competitiveness thesis), 61 percent of schools face little or no competitive pressure.  Under quite plausible assumptions, the percentage of students facing limited or no competitive pressure may be as high as 81 percent.  This, say the authors, “explains why school competition is not a policy lever to significantly improve learning and teaching.”

MythofMarkets_Figure6Why not?  First, it turns out that decisions on choice of school are made on a wide variety of bases, including distance to the school, the religious affiliation of the school, the record of the school in the sport the student is interested in playing, perceived discipline in the school, the perception that the school is safe for the student and other factors.  The report includes a scattergram plotting average school scores on the national achievement tests on one axis against increases in school enrollments on the other axis.  If parents made their choices based on student achievement, one would expect to see increases in enrollment at the schools that performed best and decreases for those schools that scored least well.  But there is in fact hardly any correlation at all between the choices made by parents and schools’ scores on the national achievement tests.

But the story does not end there.  It turns out that most parents are not prepared to have their children at a school that is not close to their home.  They are not able or willing to pay very much more to put their child in a high performing school than a lower performing one.  A large fraction are not willing to abandon schools that are, by any measure, low performing.  In these ways and many more, the market for schools does not look much like the market for motorcars or TV sets.  That turns out to be true even in Australia, which has one of the most robust systems in the world for reporting on student performance, school by school, to the parents of Australian children.  The problem there is not that parents do not have enough information to select high performing schools over lower performing schools.  There is plenty of information.  The problem is that they either will not or cannot act on it.

Jensen points out that there are other respects in which the school market is a poor imitation of a well-functioning market for goods services.  Not least is the matter of price.  In a well-functioning classic market, prices are used to clear the markets.  But governments cannot do that in education.  Education is thought to be a public good, and governments are obligated to provide it to citizens’ children irrespective of their ability to pay for it.  So, in this case, prices cannot be used to clear the market, a fact that severely compromises the very idea of a market in this arena.

In theory, the government could change some of the incentives that make this such a poor market.  One of the major sources of market failure here is the fact that there are so few places open in highly desirable schools.  Government could build extra capacity in those schools and see who comes to fill the extra spaces.  But building so much extra capacity would be highly expensive and would make the schools even less efficient—more costly per student served—than they are now, thereby defeating the purpose of introducing market forces in the first place.  In theory, government could shut down the low-performing schools, thereby driving parents to higher-performing schools.  But wherever in the free world government has tried this, it has had to back off in the face of determined public resistance.  Jensen goes through a list of such possibilities for changing the incentives to favor the development of better-functioning markets in education, and shows why they won’t work.

Jensen concludes that relying on markets is not a viable way to improve school system performance.  In the real world, few schools face significant competition.  Even when parents can easily get good information about school performance, they rarely use that information to choose higher-performing schools.  As a practical matter, there is very little that government can do to change these facts.

I do not expect this report to dent Americans’ belief that charters and choice are among the most effective levers for improving the performance of their schools.  But I thought you might be interested in knowing that there is no evidence to support that belief.