The Shape of Corruption: Colombia as a Case Study Laura Langbein American University Pablo Sanabria Universidad de Los Andes
The research question Is corruption stable? Who cares? Equilibria are hard to change, regardless of whether they are socially productive or costly Levels of corruption (high/medium/low) are likely to vary within countries Policy implication: if corruption is characterized by stable equilibria in specific places: Exogenous policy efforts to reduce corruption are likely not to be fruitful; Policies that might be fruitful in one place won t work in another
Background Corruption = use of public office for private gain Vicious social problem: no player has incentive to squeal; requires collective third party enforcement of rule of law, which is undersupplied when corruption is the norm. (Supplier wants the extra payment, buyer wants the service, no one wins by squealing) Examples: hospital, police, schools, courts, permits
Corruption Corruption is a cooperative (predatory) game Resisting corruption is a cooperative (pro-social) game Probability of defecting from pro-social choice depends on repeated interactions, and the behavior of others with whom the player interacts Implication: corrupt play is more likely in some contexts than others, and it is likely to be a sticky equilibrium: corruption (t-1) -->corruption (t)
Our study Focus: Low level corruption in cities in Colombia Stable unitary democracy with thriving legal and illegal markets (7 on Polity scale=> fragile ) Has made efforts to modernize laws in central and local governments (GI index) Do city officials ask citizens who seek service for a bribe? Individual level data over time in 55 cities (2004-2011)
Hypotheses Corruption is stable within cities Corruption is not increasing or decreasing over time Corruption is characterized by dual or multiple equilibriums unique to each city
Data: Dependent Variable LAPOP, 2004-2011 Sum of response to following questions 1) Did any public official ask you for a bribe in the last 12 months? 2) In order to process something in the municipality, a permit for instance, during the last year, has someone asked you to pay a bribe? 3) Have you had to pay a bribe at the courts during the past year? 4) Did any police officer ask you for a bribe in the last 12 months? 5) During the last 12 months, have you had to pay a bribe in order to be assisted in a hospital? 6) During the last 12 months, have you had to pay a bribe in a school? 7) At work, have you been requested to pay a bribe during the last 12 months?
Data: Independent Variables Year of survey = time trend (we expect none) Exposure: did respondent use service? (1 = yes; 0 = no) Response choice: did respondent answer the exposure/bribe request question (1=yes; 0 = no) Dummy variable for each city (Medellin = reference) (we expect only the dummies to matter)
Descriptive statistics Bribe frequencies are about 12.5% each year, no clear trend Most common bribe request: police (5.7%), municipality (5.1%) Number of bribes: most respondents score 0 (no bribes, no exposure, no response) 8% report 1 bribe, 2% report 2 bribes; range of bribe requests is from 0 to 5 Highest rate of bribes: some cities on Caribbean coast (Barranquila, Cartagena); cities in regions associated with illegal activities (Monteria, Valledupar, Santa Marta); same for small cities
Results: Zero-Inflated Poisson Count data; mean = variance Zero-inflation to account for real and no exposure 0 Standard errors clustered by city Inflation adjustment statistically (not substantively) significant No trend 42 of the 54 dummies are significant 33 of the incident rate ratios (IRR) are lower than Medellin; 9 are higher. The range is from about 1.5:1 more bribes than Medellin to about.5:1 fewer than Medellin
Results for each city No trend in most cities => equilibrium Downward trend in Bogota, despite high profile scandals (Maybe high cost/high value bribes crowd out low-level street level bribes?) (also down in 3 smaller cities) Upward trend in Cartagena (and 2 smaller cities)
Conclusions Corruption is stable in most locations in Colombia Corruption is contextual: different equilibria in different places Importance of studying corruption over time within countries
Discussion Good news if cooperative production is stable; but corruption is cooperative predation=> Dismal news No evidence in this study of sources of corruption Another concurrent paper => change in civil service laws (independent/merit-based bureaucracy) has no impact Future research: norms and education