The Political Power of Civil Society

Abstract We propose a model to understand how civil society affects political reform. A member of civil society who is a political actor can defect to the opposition and directly erode the ruler's power. An ordinary citizen, however, can only indirectly depose the ruler through popular discontent, generating governance costs which, if sufficiently large, can trigger defection to the opposition by political actors. A key factor that determines the success of reform is the degree of organization of civil society into particular groups. Wide-reaching groups whose membership spans political actors and ordinary citizens are most effective; groups composed of solely political actors are somewhat effective; while groups composed solely of ordinary citizens are the least effective.

An important distinction

Two ways to change a regime, but only one of them works alone

A ruler falls when the people whose support she needs withdraw it. So the first question is not how angry a population is, but who among them can actually withdraw support. The paper splits civil society into two components.

Their instrument is defection. The ruler needs the support of a coalition C ⊂ P; if its members switch to a challenger, the ruler is out. To prevent that, she must pay them — private transfers that a challenger is always willing to outbid.

Membership scales with regime type rather than being fixed by it. In an electoral democracy most adults are political actors. In an autocracy the set shrinks to those whose defection actually costs the ruler something.

In 1688 England: MPs, peers, bishops sitting in the Lords, military officers, judges, lords lieutenant. In 1986 Manila: the defence minister, the armed forces vice chief of staff, the bishops. In a Gulf monarchy: a few hundred people.

Their instrument is protest, and protest does not remove rulers. What it does is make patronage expensive: with μ people in the streets, every pound the ruler moves to a coalition member costs her κ(μ) + ε instead of ε. Repression, disorder and lost revenue are all in that wedge.

Protest is a strategic complement — the cost of joining falls as more join, and falls further if the joiners are in your group. That is why organization, not grievance, is the binding constraint.

In 1688 England: pamphleteers, London merchants, coffee-house patrons, petition signers, the lower clergy, common soldiers. In 1986 Manila: the parish networks that filled EDSA. In 2011 Cairo: Tahrir Square.

The two are not symmetric. Defection works with nobody in the streets: even at μ = 0, transfers still cost the ruler ε per unit, so the coalition still has to be bought. Protest with nobody to defect does nothing at all: ∂S/∂μ = 0 when |C| = 0. Popular power is real, but it is leveraged through political actors — never a substitute for them.

Interactive · Propositions 4–6

The ruler's ledger

Staying in power has a price: the minimum spending S needed to keep every coalition member from defecting. Rule survives while revenue covers it. Move the sliders, or load a case and watch which margin gave way.

Revenue at her disposal τyN + R

Price of holding the coalition S

Protest size μ

Cost of a transfer κ(μ)+ε

Paid per member R̄

A stylised illustration of the model's comparative statics, not the paper's solved equilibrium. The functional forms here are chosen to be transparent: S = ḡ + (κ(μ)+ε)·|C|·R̄, with protest μ rising in organisation and κ concave and saturating. The signs, not the numbers, are what the paper delivers.

Out of sample

Two more tests

If the argument is right, it should travel. The Philippines in 1986 supplies a near-textbook cross-cutting organisation. The Arab Spring supplies its absence.

Manila · February 1986

One organisation that contained everybody

The Catholic Church in the Philippines is close to the model's limiting case: a single group spanning almost the whole of P and O at once. Bishops sat among the political actors; parish clergy and Radio Veritas could move millions of ordinary citizens. Cardinal Sin could do both things at the same time.

Marcos was simultaneously running out of room to pay. The 1981 debt crisis, the cut-off of international lending after Ninoy Aquino's assassination in 1983, and increasingly conditional US backing all tightened his constraint — while his attempt to hold the military, business elites and local bosses together widened the coalition he had to buy. When Enrile and Ramos defected, the arithmetic was already lost.

200+civil society organisations in the KOMPIL coalition, 1984

4 daysfrom the first defections to Marcos in exile

Corazon Aquino's inauguration, February 1986 — Inauguration of Corazon Aquino, 25 February 1986. Malacañang Palace archives. Public domain, via Wikimedia Commons. Source

Crowds in Tahrir Square, February 2011 — Tahrir Square, 10 February 2011. Photograph by Jonathan Rashad, CC BY 2.0, via Wikimedia Commons. Source

The Middle East · 2011

Protest without anyone to defect

The Arab Spring was not short of protesters, and the model does not say it should have failed for lack of ideology. It failed where the two ingredients the model requires were missing: organisations that reach into the set of political actors, and rulers whose revenues could actually be squeezed.

Timur Kuran's account supplies the long-run reason. Centuries in which the was the dominant organisational form, combined with an inheritance system that fragmented capital and a legal system without corporate personhood, left few durable, autonomous organisations. Registered NGOs existed in the thousands — in Tunisia over 5,000 by the 1990s — but they were licensed by interior ministries and required to stay out of politics.

Where oil paid for loyalty, the regimes barely moved. Where it did not, rulers fell: Ben Ali went once the broke with him; Mubarak went as US support was withdrawn. But Egypt shows the sharper point: the Muslim Brotherhood was a mass organisation with no purchase on the military, and the liberals had no way to organise working people. A ruler fell; the reform did not hold.

What the model adds

Three things to take away

Composition, not size

A movement's headcount matters far less than whether its organisations contain people who can defect. The same crowd is decisive in one architecture and futile in another.

Money buys time

External rents — French subsidies, oil, a superpower patron — relax the constraint that popular pressure tightens. Cutting them off is often what converts protest into turnover.

Organisation is contingent

Nothing guarantees cross-cutting groups exist when they are needed. England's parties were an accident of the Exclusion Crisis; their absence elsewhere has roots centuries deep.

The model is deliberately narrow. It explains how a coalition unravels, not everything about any revolution — and the authors are explicit that it is not a complete account of 1688. Repression technology, the kinds of rents available to a regime, and foreign support are all flagged as open ground.

Cite this paper Desierto, Desiree A., and Mark Koyama. “The Political Power of Civil Society.” Journal of Comparative Economics, forthcoming. JEL: D72, D74, H41, N13 · Supported by Grant 62593 from the John Templeton Foundation and by an Emergent Ventures Grant.