Philip Roessler’s attentive review of my book concludes with questions about the scope of the theory’s applicability, the possibility for vote brokering in presidential elections, and the implications of traditional chiefs for state building. Although these questions will only be answered satisfactorily once additional scholarship has accumulated on traditional chiefs, I use my response to outline what my theoretical expectations are and how they diverge from Roessler’s hypotheses.

The Paradox of Traditional Chiefs in Democratic Africa emphasizes the role that chiefs play in brokering local development projects and the motivations this gives citizens for preferring electoral candidates with stronger relationships to chiefs: The rationale is that more resources are likely to flow to local communities through this brokered relationship if the connections between chiefs and politicians are stronger. The theory depends on the existence of locally embedded chiefs with some influence over the day-to-day governance of their communities. In addition, it only gives citizens motivations for supporting one candidate over another in cases in which the local chief is differentially connected to the competing candidates. I would expect the development broker model to apply in all instances in which these conditions hold. These include some regions of Francophone Africa (on this theme, see Lauren Honig, “Selecting the State or Choosing the Chief? The Political Determinants of Smallholder Land Titling,” World Development 100, 2017) and some chiefdoms in presidential elections.

Importantly, I disagree with Roessler’s hypothesis that chiefs are likely to serve as traditional vote brokers in presidential elections, relying on vote buying, coercion, and social sanctioning to mobilize citizens. Instead, my hypothesis is that the development broker model will operate, but with regional-level brokers. Voters will consider the strength of the connections between major presidential candidates and regional leaders (including some higher-level chiefs), because these leaders are the most relevant brokers of resources from the president. Thus, in both parliamentary and presidential elections, I hypothesize that voters are primarily concerned with evaluating how well politicians will perform in delivering to their communities if elected.

Roessler also raises important questions about the relationship between the strength of chiefs and the strength of the formal bureaucracy. In particular, he raises concerns that politicians might be discouraged from investing in the bureaucracy in cases in which they can respond to citizens through traditional institutions. Does this in turn harm the ability of the state to provide public goods that must be provided on a wide geographic scale? Would investments in the bureaucracy eventually result in a fairer allocation of local public goods? These questions follow naturally from the framework in Roessler’s book, which emphasizes the inherent instability of regimes dependent on informal alliances with ethnic big men.

In contrast, I am both more sanguine about the types of governance that result from alliances with traditional chiefs and more pessimistic about the types of politics that would emerge if politicians invested directly in the formal bureaucratic apparatus. Strong local chiefs are very effective substitutes for the formal bureaucracy when it comes to delivering local public goods. This helps solve the “democratization backwards” problem because it allows elected officials to meet the highest-priority demands of rural citizens, even without a formal bureaucratic apparatus to assist them.

Furthermore, I am skeptical that weakening traditional chiefs would lead to better delivery of either national or local public goods in the medium term. In the absence of traditional leaders, I would not expect states to invest more in bureaucratic agencies that provide national public goods. In fact, the ability to deliver local public goods through traditional leaders—a cheaper substitute apparatus—frees up funds for the army, the electrification authority, and the anti-corruption commission. If there is resistance to investing in these agencies, I do not think it is coming from local traditional chiefs.

Without chiefs, I would expect states to invest more in the arms of the bureaucracy that deliver local public goods like schools, clinics, and wells. But, unfortunately, I would not expect the bureaucratic apparatus to be fairer in its allocation of these goods. Partisan bureaucracies are widespread in developing democracies (see Allen Hicken, “Clientelism,” Annual Review of Political Science 14, 2011). A strengthened bureaucratic apparatus would almost certainly continue to favor certain citizens over others, creating a distribution of resources that may even be less equal than when politicians must work with traditional leaders to provided targeted goods.