%%{init: {"theme":"base","width":1150,"height":620,"htmlLabels":false,"flowchart":{"useMaxWidth":true,"htmlLabels":false,"nodeSpacing":70,"rankSpacing":130,"padding":18,"curve":"basis"},"themeVariables":{"fontSize":"22px","fontFamily":"DM Sans, Segoe UI, sans-serif","primaryColor":"#e6ecf2","primaryTextColor":"#1e293b","primaryBorderColor":"#334155","lineColor":"#334155","edgeLabelBackground":"#f9fafb","tertiaryTextColor":"#1e293b"}}}%%
flowchart LR
P["Party wants<br/>your vote"] --> A["PROGRAMMATIC<br/>promise a rule"]
P --> B["CLIENTELIST<br/>hand over a thing"]
A --> A2["A pension for<br/>everyone over 65"] --> A3["You get it<br/>either way"]
B --> B2["A food parcel<br/>for you, this week"] --> B3["Only if<br/>you comply"]
classDef prog fill:#dcebe4,stroke:#4a7c6f,stroke-width:2.5px,color:#1e293b;
classDef clie fill:#f7e2da,stroke:#b44527,stroke-width:2.5px,color:#1e293b;
classDef root fill:#e6ecf2,stroke:#334155,stroke-width:2.5px,color:#1e293b;
class A,A2,A3 prog; class B,B2,B3 clie; class P root;
The Politics of Corruption
Lecture 9: Clientelism and Vote Buying
Today
One question runs through the whole hour:
If the ballot is secret, why pay for a vote you cannot verify?
By the end, you should be able to:
- Distinguish clientelism, patronage, vote buying, and programmatic politics
- Explain how machines monitor voters despite the secret ballot
- Explain why buying turnout sidesteps the monitoring problem
- Explain reciprocity as an enforcement device needing no surveillance
A Tuesday before the election
- A woman knocks. You know her: she lives three streets away.
- She leaves a food parcel. No threat, no contract, no receipt.
- She mentions, lightly, that she will be at the polling station.
- On Sunday you enter a booth. Alone. Curtain closed.
- Nobody can see your ballot. So why did she bother?
Hold on to her. We come back to her at the end.
This is not rare
And those are floors, not ceilings
The puzzle in one line
If the ballot is secret, why pay for a vote you cannot verify?
- The voter can take the food and vote as they please.
- Parties spend real money on this across very different electoral systems.
- Either parties are foolish, or we have the mechanism wrong.
I. What Exchange Are We Talking About?
Thomas Nast, “The Tammany Tiger Loose,” Harper’s Weekly, 1871 — the New York machine as a beast in the arena. Public domain, via Wikimedia Commons.
Two logics of winning votes
Author’s illustration, following Hicken (2011).
Hicken’s four elements
A relationship is clientelist when all four hold:
- Dyadic — a two-sided relationship between specific people
- Contingent — delivery depends on what the citizen does
- Hierarchical — the patron holds more power than the client
- Iterated — it repeats; both sides expect a next time
Keep these straight. Clientelism is the relationship; vote buying is one of its electoral forms. Scholars differ on whether a one-off pre-election handout still counts — Finan and Schechter deliberately use “vote-buying” broadly, to include clientelism.
Source: Hicken (2011, pp. 291–293); definitional note from Finan & Schechter (2012, p. 1, n. 4).
Two questions, four cells
| Non-contingent | Contingent on behaviour | |
|---|---|---|
| Universal benefits | A national pension | Rare, hard to sustain |
| Particularistic benefits | Pork; a one-off gift | Clientelism |
Clientelism sits in one cell: particularistic and contingent.
Author’s illustration, adapted from Hicken (2011).
Five terms, one example each
| Term | Who receives | What is exchanged | Example |
|---|---|---|---|
| Programmatic politics | Any qualifying citizen | Policy rules | Universal pension law |
| Patronage | Party insiders | Public jobs, contracts | Payroll posts after victory |
| Vote buying | Ordinary voters | Cash, food, goods | Parcel before election day |
| Turnout buying | Own passive supporters | Cash, transport, favours | A ride to the polls |
| Clientelist service | Neighbourhood residents | Access to the state | Getting your water reconnected |
Author’s illustration, drawing on Hicken (2011) and Nichter (2008).
The running case: Mexico’s PRI
- Founded 1929; held the presidency without interruption until 2000
- Organised society into sectors: labour (CTM), peasants (CNC), popular (CNOP)
- Benefits — land, jobs, credit, relief — flowed through these bodies
- Membership and loyalty were the price of access
- In 2000 the PRI candidate took 36.1%, and lost the presidency
A machine is not a conspiracy. It is an organisation for distributing things. We will keep asking of the PRI: what exactly was it buying?
Sources: Magaloni (2006); Díaz-Cayeros, Estévez & Magaloni (2016). 2000 vote share reported in Klesner (2001).
The machine needs a middleman
%%{init: {"theme":"base","width":1150,"height":580,"htmlLabels":false,"flowchart":{"useMaxWidth":true,"htmlLabels":false,"nodeSpacing":75,"rankSpacing":140,"padding":18,"curve":"basis"},"themeVariables":{"fontSize":"22px","fontFamily":"DM Sans, Segoe UI, sans-serif","primaryColor":"#e6ecf2","primaryTextColor":"#1e293b","primaryBorderColor":"#334155","lineColor":"#334155","edgeLabelBackground":"#f9fafb","tertiaryTextColor":"#1e293b"}}}%%
flowchart LR
L["Party<br/>leaders"] -->|"resources"| B["Broker<br/>local and long-standing<br/>knows who is loyal<br/>and who is buyable"]
B -->|"goods and favours"| V["Voters in the<br/>neighbourhood"]
V -->|"votes and turnout"| L
classDef party fill:#e6ecf2,stroke:#334155,stroke-width:2.5px,color:#1e293b;
classDef brok fill:#f7e2da,stroke:#b44527,stroke-width:3px,color:#1e293b;
classDef vote fill:#dcebe4,stroke:#4a7c6f,stroke-width:2.5px,color:#1e293b;
class L party; class B brok; class V vote;
In PRI-era Mexico the middleman usually sat inside a sectoral organisation — a union local or an ejido committee. Remember this box: every theory today runs through it.
Author’s illustration, following Stokes (2005) and Stokes, Dunning, Nazareno & Brusco (2013).
Exercise 1 — Sort the transaction
In pairs (5 minutes). Give each case one of the five terms from the table.
- A law raises the pension for everyone over 65.
- The mayor puts 30 campaign workers on the payroll.
- An operative hands out food parcels before the vote.
- A local leader fixes your water — and keeps a list.
- A party van collects known supporters on election morning.
Then: which two were hardest to separate, and why?
So what?
- We can now say what clientelism is: particularistic and contingent.
- But “contingent” is a promise about something the party cannot see.
- Secret ballots were invented to make vote choice invisible.
- So the definition contains a puzzle it cannot solve alone.
Next: how can a contingent exchange survive a secret ballot?
II. The Secret Ballot Problem
What the secret ballot changed
Before. Parties printed their own ballots — distinctive paper, visible in the hand.
After. One official ballot, marked behind a screen: el voto es secreto.
Massachusetts 1888; Argentina’s Sáenz Peña Law 1912. Compliance stops being visible.
Left: Cornell University Library (1888). Right: ProtoplasmaKid, CC BY-SA 4.0 (2018). Both via Wikimedia Commons; full credits at the end.
Stokes: accountability, inverted
%%{init: {"theme":"base","width":1150,"height":600,"htmlLabels":false,"flowchart":{"useMaxWidth":true,"htmlLabels":false,"nodeSpacing":60,"rankSpacing":110,"padding":16,"curve":"basis"},"themeVariables":{"fontSize":"21px","fontFamily":"DM Sans, Segoe UI, sans-serif","primaryColor":"#e6ecf2","primaryTextColor":"#1e293b","primaryBorderColor":"#334155","lineColor":"#334155","edgeLabelBackground":"#f9fafb","tertiaryTextColor":"#1e293b"}}}%%
flowchart LR
H1["NORMAL"] --> N1["Party acts<br/>in office"] --> N2["Voter judges<br/>the record"] --> N3["Voter punishes<br/>at the ballot box"]
H2["PERVERSE"] --> P1["Voter acts<br/>in the booth"] --> P2["Party infers<br/>the vote"] --> P3["Party punishes:<br/>goods cut off"]
classDef norm fill:#dcebe4,stroke:#4a7c6f,stroke-width:2.5px,color:#1e293b;
classDef perv fill:#f7e2da,stroke:#b44527,stroke-width:2.5px,color:#1e293b;
classDef lab fill:#1e293b,stroke:#1e293b,stroke-width:2px,color:#f9fafb;
class N1,N2,N3 norm; class P1,P2,P3 perv; class H1,H2 lab;
Author’s illustration of the central claim in Stokes (2005).
How the threat is sustained
Stokes models machine and voter as a repeated game with imperfect monitoring.
- It cannot read your ballot, but it reads your street.
- Brokers live where you live: they see rallies, gossip, signals.
- From those signals the machine forms a guess about your vote.
- A guess is enough: it only needs to be right often.
- What disciplines you is losing the flow, not one parcel.
- The exchange is enforced by the future, not the transaction.
Source: Stokes (2005, pp. 316–320).
Predictions, and what Argentina showed
The model predicts
- Poverty. Poorer voters are cheaper to buy.
- Monitoring. More observable places, more vote buying.
- Ideology. Skip the committed on both sides.
The survey data show
- Poorer respondents far more likely to report targeting.
- More vote buying in smaller communities.
- Strong Peronists less likely to be bought.
Many respondents believed parties could find out how they voted. That belief does the work — true or not.
Sources: Stokes (2005, pp. 318–322); Brusco, Nazareno & Stokes (2004). Cross-nationally, wealth is negatively associated with being targeted across 22 countries (Faughnan & Zechmeister, 2011).
Exercise 2 — The broker’s list
In pairs (5 minutes). You are a broker. You hold 10 food parcels and a list of 40 households.
- Which households get a parcel, and why those?
- How would you find out whether they kept their side?
- What changes if you can see whether they voted, but never how?
Be ready to say what information you wished you had.
So what?
- Stokes solves the puzzle by making monitoring imperfect but real.
- Yet in most places brokers cannot actually observe the ballot.
- And question 3 pointed somewhere else entirely.
- Turnout is often recorded and verifiable. Vote choice is not.
Next: two ways the exchange works without watching the ballot.
III. Turnout, Reciprocity, and Enforcement
A lever voting machine, Old State House, Hartford, Connecticut. Photo by Kenneth C. Zirkel, CC0, via Wikimedia Commons.
Nichter: buy turnout, not votes
- Suppose the party pays its own supporters who usually stay home.
- It is not asking them to switch, only to appear.
- Turnout is recorded, observable, and often publicly verifiable.
- Enforcement then no longer needs the ballot at all.
- Argentine data reanalysed: rewards flowed to strong Peronists.
Ask it of our running case: was the PRI buying votes, or buying turnout?
Source: Nichter (2008, pp. 19–20, 26–29).
Who gets targeted? Two theories
| Voter type | Likely to vote anyway | Unlikely to vote |
|---|---|---|
| Strong supporter | Waste of money | Nichter: buy turnout |
| Weak opponent | Stokes: buy the vote | Two things to change |
- Stokes must monitor how you vote; Nichter only whether.
- The theories predict opposite targets — so data can adjudicate.
Author’s illustration, after Stokes (2005) and Nichter (2008).
Finan & Schechter: reciprocity
Rural Paraguay: a household survey after the 2006 municipal elections, plus a survey of the brokers themselves.
- 26% of households said a party offered them money, food, or goods.
- Reciprocity was measured years earlier, in an anonymous trust game.
- Anonymous play is a disposition, not a favour to anyone.
- A one-standard-deviation rise in reciprocity raised targeting by 44%.
- Targeted villagers were not closer to the broker in the network.
Why this answers our puzzle. The party never has to observe your ballot. If the obligation lives inside the voter, nothing external needs enforcing.
Source: Finan & Schechter (2012, pp. 863–864, 869–875).
But can brokers actually find them?
| What the broker was asked about a villager | Accuracy |
|---|---|
| Years of schooling | Correlation 0.73 |
| Name of the spouse | 84% correct |
| Whether the person voted in 2006 | 71% correct (56% by guessing) |
| Which party the person supports | 77–83% correct |
Identical twins report each other’s schooling at 0.87–0.92. Brokers are not far behind.
Source: Finan & Schechter (2012, Table II, pp. 870–872). Twin comparison: Ashenfelter & Krueger (1994), cited therein.
Three answers, one parcel
%%{init: {"theme":"base","width":1150,"height":480,"htmlLabels":false,"flowchart":{"useMaxWidth":true,"htmlLabels":false,"nodeSpacing":45,"rankSpacing":150,"padding":14,"curve":"basis"},"themeVariables":{"fontSize":"20px","fontFamily":"DM Sans, Segoe UI, sans-serif","primaryColor":"#e6ecf2","primaryTextColor":"#1e293b","primaryBorderColor":"#334155","lineColor":"#334155","edgeLabelBackground":"#f9fafb","tertiaryTextColor":"#1e293b"}}}%%
flowchart LR
G["A parcel<br/>arrives"] --> S["STOKES<br/>she infers your vote"]
G --> N["NICHTER<br/>she sees you turned out"]
G --> F["FINAN AND SCHECHTER<br/>you feel obliged"]
S --> V["The machine gets<br/>what it paid for"]
N --> V
F --> V
classDef gift fill:#e6ecf2,stroke:#334155,stroke-width:2.5px,color:#1e293b;
classDef mech fill:#f7e2da,stroke:#b44527,stroke-width:2.5px,color:#1e293b;
classDef out fill:#dcebe4,stroke:#4a7c6f,stroke-width:2.5px,color:#1e293b;
class G gift; class S,N,F mech; class V out;
The broker makes all three possible.
Author’s illustration, synthesising Stokes (2005), Nichter (2008) and Finan & Schechter (2012).
IV. Why It Survives
Why development does not end it
%%{init: {"theme":"base","width":1150,"height":460,"htmlLabels":false,"flowchart":{"useMaxWidth":true,"htmlLabels":false,"nodeSpacing":70,"rankSpacing":120,"padding":18,"curve":"basis"},"themeVariables":{"fontSize":"21px","fontFamily":"DM Sans, Segoe UI, sans-serif","primaryColor":"#e6ecf2","primaryTextColor":"#1e293b","primaryBorderColor":"#334155","lineColor":"#334155","edgeLabelBackground":"#f9fafb","tertiaryTextColor":"#1e293b"}}}%%
flowchart LR
A["Poverty makes<br/>a vote cheap"] --> B["Machine buys<br/>votes and turnout"]
B --> C["It underinvests in<br/>public goods, which<br/>cannot be switched off"]
C --> A
classDef loop fill:#f7e2da,stroke:#b44527,stroke-width:2.5px,color:#1e293b;
class A,B,C loop;
Clientelism is not a stage countries outgrow — it is an equilibrium that reproduces its own conditions. Competition can even intensify it: close races raise the value of each bought vote. In a Benin field experiment, clientelist appeals outperformed programmatic ones.
Author’s illustration, following Hicken (2011, pp. 296–302). Benin experiment: Wantchekon (2003).
Back to Tuesday
She gave you food on Tuesday. She could not see your ballot on Sunday. Why bother?
- Stokes: she thinks she can infer what you did — and cut you off.
- Nichter: she may only care that you showed up at all.
- Finan & Schechter: she chose you because you are the type who repays.
Same parcel. Three different political mechanisms — and accountability running backwards: the voter answers to the party.
Session 12 takes the same broker logic into organised crime, where the good sold is protection.
References
References
Ashenfelter, O., & Krueger, A. (1994). Estimates of the economic return to schooling from a new sample of twins. American Economic Review, 84(5), 1157–1173.
Brusco, V., Nazareno, M., & Stokes, S. C. (2004). Vote buying in Argentina. Latin American Research Review, 39(2), 66–88.
Díaz-Cayeros, A., Estévez, F., & Magaloni, B. (2016). The political logic of poverty relief: Electoral strategies and social policy in Mexico. Cambridge University Press.
Faughnan, B. M., & Zechmeister, E. J. (2011). Vote buying in the Americas (AmericasBarometer Insights No. 57). Latin American Public Opinion Project, Vanderbilt University.
Finan, F., & Schechter, L. (2012). Vote-buying and reciprocity. Econometrica, 80(2), 863–881.
Gans-Morse, J., Mazzuca, S., & Nichter, S. (2014). Varieties of clientelism: Machine politics during elections. American Journal of Political Science, 58(2), 415–432.
González-Ocantos, E., Kiewiet de Jonge, C., Meléndez, C., Osorio, J., & Nickerson, D. W. (2012). Vote buying and social desirability bias: Experimental evidence from Nicaragua. American Journal of Political Science, 56(1), 202–217.
Hicken, A. (2011). Clientelism. Annual Review of Political Science, 14, 289–310.
References (continued)
Klesner, J. L. (2001). The end of Mexico’s one-party regime. PS: Political Science & Politics, 34(1), 107–114.
Magaloni, B. (2006). Voting for autocracy: Hegemonic party survival and its demise in Mexico. Cambridge University Press.
Nichter, S. (2008). Vote buying or turnout buying? Machine politics and the secret ballot. American Political Science Review, 102(1), 19–31.
Stokes, S. C. (2005). Perverse accountability: A formal model of machine politics with evidence from Argentina. American Political Science Review, 99(3), 315–325.
Stokes, S. C., Dunning, T., Nazareno, M., & Brusco, V. (2013). Brokers, voters, and clientelism: The puzzle of distributive politics. Cambridge University Press.
Wantchekon, L. (2003). Clientelism and voting behavior: Evidence from a field experiment in Benin. World Politics, 55(3), 399–422.
Popescu (TEC) Politics of Corruption Lecture 9