Why is learning in primary education so poor, and how can it be improved? This remains a central question for developing countries today. One explanation focuses on the role of politicians—an approach followed by Agustina Paglayan’s research presented at this year’s RISE conference.

There is a good reason to look at the role of politicians: expert advice—even the best, based on rigorous research—is just one input to the decision making of politicians, who may care about a range of interests from different groups, as well as their own self-interest. 

In her presentation, Paglayan zooms in on three potential stakeholder groups: teacher unions, citizens electorate, and politicians themselves. Her conclusion is bold: autocratic politicians’ own interests have historically played a central role in shaping primary education systems, more so than other factors. As a consequence, the resulting systems tend to be good at delivering socialisation (originally to encourage nation- and state-building); but they are poorly equipped to produce learning—indeed, they were never designed to be. 

There is a great deal to be learned from this work. But let’s first consider the findings in more detail. For those interested, the entire presentation can be found here (although the papers are discussed in a different order below, as indicated).

Teacher unions are often seen as a major driving force behind the expansion of school systems—they tend to advocate for higher number of teachers, better salaries, more school inputs, and larger school budgets. But are teacher unions successful in achieving these objectives? Paglayan’s first paper (second paper in her presentation; the paper is forthcoming in the American Journal of Political Science) investigates this for the United States, a case that has informed how we think about teacher unions, not just in the United States, but around the world. The paper examines whether granting collective bargaining rights to teachers led to an expansion of the education system, applying difference-in-differences to a new longitudinal data set covering all of the United States and spanning 90 years. She finds that it did not. In most cases the enlargement took place before collective bargaining rights were granted, which was not well accounted for in previous analyses. Moreover, many of the laws that granted these rights included both pro- and anti-union provisions. In summary: teacher unions seem to have had limited influence on the expansion of an education system in terms of student-teacher ratio, average teacher salary, or expenditure per pupil. 

A second paper (first paper in her presentation) turns to the role of the electorate. The positive relationship between democracy and education enrolment has received much attention in political science and economics. There is no dearth of theory, but rigorous empirical analysis is scarcer. Using newly compiled country-level data for the last two centuries, Paglayan finds that democratisation had little or no impact on primary school enrolment rates. Because state-controlled primary education emerged about a century before democratisation, in most countries a majority of the population [and hence the proverbial “median voter”] already had access to primary education before the regime change. Conclusion: primary education systems did not develop or expand as a response to democratic pressure from the electorate.  

A third paper looks at the role of politicians’ own interests. Building on previous work that emphasised the role of nation- and state-building for the expansion of education, Paglayan centers on the role of civil war and domestic conflict. The notion that education can foster a shared language, a common culture, and possibly deference to the state’s authority and aligned political values, is intuitively appealing, but there is scant evidence. Using three types of analysis—cross-country analysis from Latin America and Europe, within-country analysis of 19^thcentury Chile, and qualitative historical assessment of Argentina and Prussia—the paper concludes that state investment in primary education came in response to civil war, with the supposed objective to instil values that maintain order and help prevent future rebellion. Another recent paper draws complementary inference, suggesting that compulsory schooling in North America was introduced to instil civic values to the culturally diverse migrants that had just arrived (see Bandiera et al., 2018). The focus on socialisation as a key objective of primary education is common in more sociological analyses (see for instance Bowles and Gintis, 1976, 2002). In summary, education systems were designed to instil ‘the morality of the masses and obedience to the king’ as one member in the audience put it.

What do we learn from this? I see six lessons from Paglayan’s work—there may be more. 

First, politics matters. We all agree. More research is needed on how politics matters and this is why RISE has initiated a separate work stream in this area. Paglayan’s work provides great starting points on three decisive elements of education systems: Who are the stake holders, what is their role, and what processes are they involved in?  

Second, primary schools seek to socialise their pupils. They typically aim to ‘produce’ more or less well-behaved citizens. This aids to comprehend the limitations of current and future systems. The objective of socialisation may for instance help explain the dominance of the state in the provision of education, for example because socialisation is hard to contract out, as argued by Pritchett and Viarengo (2015). As a consequence, the role of alternative approaches involving the private sector is limited, but may have merit under certain conditions. More such work is needed to help understand the promises and limitations of contracting out, in particular given the rise in low-cost private schooling (see here for Pakistan); it may also inform on-going experimentation with outsourcing (see here for Liberia). 

Third, post-conflict settings may offer unrivalled opportunities to (re)build and (re)design education systems. Often characterised by an increased willingness to cooperate, these societies may be more ready to invest in public goods (to illustrate that these opportunities arise continuously: the International Crises Group identified 2 cases out of more than 70, Macedonia and Yemen, that could be categorised as transitioning out of conflict in September 2018, see here). Beyond this, the lessons become more speculative. In recent times and other settings, the relationship between civil war and education enrolment may be weaker. A growing body of micro evidence demonstrates the negative impacts of conflict on education and learning (see for instance for Rwanda (Guariso and Verpoorten, 2015) Timor Leste (Justino et al., 2013), Tajikistan (Shemyakina, 2011), Guatemala (Chamarbagwala and Moran, 2011), Cambodia (De Walque, 2004), and post-World War II Germany and Austria (Ichino and Winter-Ebmer, 2004). One interpretation that reconciles these findings is that states may need to have some degree of capacity or legitimacy to bring about investment in education. Here again, Paglayan’s findings point towards fruitful areas of research. 

Fourth, gaps exist between perception and evidence on the role of trade unions. Teacher unions can be powerful—among the most powerful in the public sector. Their activities can be hampering, neutral, or in support of student learning. Existing case study evidence mostly highlights their counteractive role, but not always [for Chile, Ecuador, Mexico, Peru, and a brief overview across Latin America and the Caribbean see Bruns and Luque, 2014; for Kenya see Bold et al., 2018; for South Africa see Zengele, 2013; for India see Kingdon and Teal, 2010]. Previous work highlights the fundamental tension between teachers’ interests (salary, employment protection, number of teachers) and pupils’ interests (learning) at the heart of a union’s purpose. While investment in complementary inputs (say books) may have higher returns than increases in teacher salary, teachers have (a) voice, books don’t. More structured analysis across settings may shed light on the circumstances under which unions facilitate or block reforms aimed at improving student learning. Paglayan’s work suggests that whether unions can make a credible threat of strike is an important determinant of their power.

Fifth, education expansion, measured in terms of enrolment, did not require democracy. Another recent study finds a similarly weak relationship between learning outcomes and democracy (Dahlum and Knutsen, 2018 using a new global data set on student learning). In neither case can we exclude that preceding economic and political institutions jointly caused expansion of education and democracy, whatever their respective timings (cf. Acemoglu and Robinson). Nevertheless, the results may also suggest that the ‘long route of accountability’ is relatively ineffective at improving either education quantity or quality—especially if these results hold for more recent periods. As it stands we have little comprehension why this is the case. Does the link between electorate and politicians fail, or do politicians’ plans fall through upon implementation? Or both? We also have limited understanding whether alternative, short routes, that put direct pressure on schools, are more effective. Emerging evidence suggests that competitive pressure and choice can have some impact (see here). Rigorous evidence on the effect of beneficiary and community involvement (voice) remains thin; on-going analysis of an intervention in Uganda suggests that it can have positive effects on student learning, pupil presence, and teacher attendance (see Barr et al., 2018).

Sixth, Paglayan’s work pushes the puzzle ahead and raises pivotal questions: Why do some education systems deliver learning while others do not? Do politicians in these contexts have alternative aims? Are they more susceptible to other pressures or interests? Does demand for welfare and economic performance play a role? Does pressure from employers and industry matter? Or did other factors create a more conducive environment? These questions will undoubtedly be the subject of future inquiry.