Bob: Thanks for the comment, which permits me to clarify a few issues that I hope will take the discussion forward.
First, on teachers and teachers unions, I was referring to their protests and objections to attempts at increasing accountability, most of which were aimed at the ultimate goal of improving learning outcomes. In addition to the examples in Merilee Grindle’s book, I would mention RECURSO in Peru where 2nd graders were sent home with a 60-word paragraph that their parents could test whether they could read in a minute. The program was a way for parents to know whether their children were learning (and, implicitly, whether the teacher was doing a good job). Teachers objected to the program on grounds that there was more to education than reading. Examples in low-income settings are equally common, as Tara Beteille’s work (and Geeta Kingdon and Mohammed Muzamil’s earlier work) on India shows.
Turning to Africa, the country with the strongest teachers union is South Africa, where there is at least anecdotal evidence that SADTU has resisted attempts at making teachers more accountable. And South Africa has seen one of the largest increases in public spending on education with almost no improvement in learning outcomes. Finally, a recent randomized control trial to introduce contract teachers in Kenya is telling. They introduced the program in two ways: one administered by an NGO, the other by the government. In the NGO-run program, there was a statistically significant improvement in learning outcomes; in the government administered program, there was none. The authors attribute the difference to the protest waged by the Kenyan National Union of Teachers that led to the government’s altering the program’s design, which affected implementation.
I mention these cases where teachers and teachers unions resisted attempts at increasing accountability to point out that teachers and politicians (and in some cases, government officials) are caught in a system of low accountability and, correspondingly, weak learning outcomes. To improve learning outcomes, we need to strengthen teachers’ accountability to the children (or their parents). This is not easy. Absentee teachers are earning a rent: they are getting paid for a job even if they don’t show up. No one likes to see their rents dissipated. But I doubt we can improve student learning outcomes without strengthening teacher accountability. Once we recognize this, we need to work with teachers to develop a new system with clear lines of accountability and monitoring of outcomes such as student learning.
Secondly, my proposal to have only a learning goal (which, incidentally, is not a new idea) is based on a notion that has a long tradition in economics, not to mention common sense. If you have two goals, one of which is easier to achieve (or at least monitor) than the other, there is a disproportionate amount of resources devoted to the easier goal. This is the problem with having both enrolment and learning as goals: the former is much more straightforward to achieve, which is why we see nearly 100 percent enrolment rates with below-50-percent learning outcomes. I was suggesting that, if we had one goal—100 percent learning outcomes (basic reading, writing and numeracy of all primary-school-age children)—that would focus all stakeholders on this goal. This goal applies to the whole cohort. It would not be abandoning the out-of-school children. On the contrary, it would require getting them into school, but we would not stop there; we would make sure they learned to read and write.