Voting in Clientelistic Social Networks: Evidence From the Philippines*

• Author: Nico Ravanilla,Michael Davidson,Allen Hicken
• Published date: 01 September 2022
• Date: 01 September 2022
• DOI: http://doi.org/10.1177/00104140211060275
• Subject Matter: Articles

Article

Comparative Political Studies

2022, Vol. 55(10) 1663–1697

© The Author(s) 2022

Article reuse guidelines:

sagepub.com/journals-permissions

DOI: 10.1177/00104140211060275

journals.sagepub.com/home/cps

Voting in Clientelistic

Social Networks:

Evidence From the

Philippines*

Nico Ravanilla, Michael Davidson Jr., and Allen Hicken

Abstract

In clientelistic environments, voters want to know which politicians are most

likely to deliver on targeted benef‌its. We argue that in these contexts, voters

use their social proximity with candidates as heuristics to inform vote choice.

To test our theory, we rely on local naming conventions to reconstruct family

networks spanning one whole city in the Philippines and assess blood and

marriage links between voters and local candidates. We then collect survey

data on pre-election candidate leanings and actual voting behavior of 894

randomly drawn voters. We show that the degrees of separation between

voters and candidates explain not only aggregate electoral outcomes, as

previous studies have found, but also individual vote choice, controlling for

pre-election leanings. We demonstrate that this is because private induce-

ments are channeled through family networks. These f‌indings highlight the

electoral importance of social proximity with politicians as an information

shortcut when voters are choosing whom to support at the polls.

Keywords

social networks, vote-buying, clientelism, voter behavior, Philippines

1

University of California San Diego, La Jolla, CA, USA

2

Microsoft Corp, Redmond, WA, USA

3

University of Michigan, Ann Arbor, MI, USA

Corresponding Author:

Nico Ravanilla, University of California San Diego, 9500 Gilman Dr, La Jolla, CA 92093, USA.

Email: nravanilla@ucsd.edu

Introduction

In recent years, social networks have become central to the study of cli-

entelism, building on a broader literature concerned with how social networks

inf‌luence voter behavior.

1

In particular, scholars have explored how, in many

clientelistic contexts, social ties outside of formal political netw orks (par-

ticularly ethnic, religious, or regional ties) can shape vote choices and can-

didate strategies (Adida, 2015;Carlson, 2015;Aspinall et al., 2011;Stithorn,

2012;Carlin et al., 2015). However, the measure of social ties among voters

and candidates within such networks has typically been based on some at-

tribute common to nodes in the networks. By and large, scholars have

overlooked the importance of the social distance between voters and poli-

ticians within the relevant social network in explaining individual vote choice.

In this article, we show that the kinship ties between voters and political

candidates—the familial distance between them—help voters decide whom to

support at the polls. Consistent with recent scholarship on social networks and

clientelism,

2

we demonstrate that this is because family networks serve as

channels for targeting private inducements for votes.

Why might social proximity with politicians per se matter for voters? A

good starting point is to think about this question from the perspective of

political campaigns. In many of these clientelistic contexts, the very nature of

the political exchange necessitates reliance on preexisting social networks.

The distribution of private inducements for votes is typically achieved by

employing brokers that are deeply embedded in social networks (Stokes,

2005). Consequently, these social networks—be they of families, clans, tribes,

or ethnic groups—inform not only which politicians run for off‌ice (Cruz et al.,

2017) but also the targeting strategies that their political machines employ

(Cruz, 2019;Ravanilla et al., 2021).

Building on these existing studies, we argue that voter social ties with

politicians can provide cues as to who is most likely to deliver on promises of

clientelistic benef‌its, much in the same way that formal political networks

(e.g., political parties) in developed democracies tell voters something about

candidates’policy preferences and the credibility of their promises. The more

socially proximate the voters are to political campaigns, the stronger the

mechanisms for sustaining the clientelistic exchange of targeted inducements

for votes, and therefore, the more likely voters will support vote-buying

candidates. Yet although we have empirical evidence on the importance of

social networks for politicians and brokers, we have little systematic evidence

on the importance of social networks for voters.

Of particular relevance for the purpose of our study is the work by Cruz

et al. (2017) about the impact of family networks on candidate performance in

the Philippines. Using data on family networks across the Philippines they are

able to demonstrate that (a) political candidates are disproportionately drawn

1664 Comparative Political Studies 55(10)

from more central families and (b) family network centrality is associated with

higher vote-shares for candidates. Like our study, the logic of their argument

rests on the distance between candidates and voters—voters with closer ties to

a candidate rely on fewer intermediaries to access that candidate, and hence,

have an increased likelihood of receiving goods and services (p. 3007). Their

empirical strategy makes use of data about candidate centrality in family

networks, and the relative centrality of a candidate’s family network in a given

village. However, while their results are consistent with their theory, due to the

lack of requisite data, Cruz et al. are unable to directly estimate the effect of

social distance on voter behavior. Using a unique dataset, our study is able

draw on individual-level measures of voting intentions and behavior to es-

tablish the micro-foundations of some of Cruz et al.’sf‌indings.

At least two obstacles make it diff‌icult for scholars to provide convincing

empirical evidence that voters use social proximity with politicians as heu-

ristics, and this diff‌iculty is best understood from a networks analysis per-

spective. The f‌irst diff‌iculty, common to research on social networks and

political outcomes generally, is in accurately mapping the network. Incom-

plete social networks can bias f‌indings in a variety of important ways and

make it more diff‌icult to rule out alternative explanations, including random

clustering, homophily (selection), and contextual (or environmental) effects

(Fowler et al., 2011). The second hurdle is more specif‌ic to the outcome of

interest: vote choice. Scholars tend to observe vote choice at a single moment

in time, that is, after the election. Without controlling for baseline candidate

leanings, correlating this static outcome with network characteristics leads to

all sorts of problems endemic to observational studies, which we describe in

the theory section below.

To overcome these empirical challenges and explore the effect of family

connections on voter behavior, we measure voter aff‌inities toward candidates

before and after the period of active campaigning and vote-buying. This

allows us to separate the impacts of vote-buying from initial leanings, both of

which might be a function of connectivity. A naming convention in the

Philippines allows us to fully map local family networks from publicly

available voter rolls, providing a close approximation of actual blood and

marriage ties between voters. We then identify candidates and randomly

sample voters within these networks to examine whether voter-candidate

familial ties are important for individual vote choice.

We start by demonstrating the importance of voter family networks on

aggregate electoral outcomes. To do so, we simulate a hypothetical election

were family relations is the only basis for choosing a particular candidate. In

this hypothetical scenario, every voter chooses the candidate to whom they are

most closely related. We tally the number of individuals who would vote for

each candidate in this hypothetical election and compare these predictions to

Ravanilla et al. 1665