Data and Code for "The Real State: Inside the Congo's Traffic Police Agency"

Principal Investigator(s): View help for Principal Investigator(s) Raul Sanchez de la Sierra, University of Chicago and National Bureau of Economics Research; Kristof Titeca, Antwerp University ; Stan (Haoyang) Xie, University of Chicago; Aimable (Amani) Lameke, Marakuja Kivu Research; Albert Malukisa Nkuku, University Catholique du Congo

Version: View help for Version V1

Name	File Type	Size	Last Modified
replication_police			09/20/2024 10:11:AM

Project Citation:

Sanchez de la Sierra, Raul, Titeca, Kristof , Xie, Stan (Haoyang), Lameke, Aimable (Amani), and Nkuku, Albert Malukisa. Data and Code for “The Real State: Inside the Congo’s Traffic Police Agency.” Nashville, TN: American Economic Association [publisher], 2024. Ann Arbor, MI: Inter-university Consortium for Political and Social Research [distributor], 2024-10-23. https://doi.org/10.3886/E208442V1

Project Description

Summary: This entry is data and or code accompanying the article "The Real State: Inside the Congo's Traffic Police Agency." The summary of the study presented in the paper is as follows. This paper provides insight into a corruption scheme in Kinshasa's traffic police agency. First, various data collection branches show that the agency's revenue is five times that from fines and is derived from a coalition of traffic police officials, their managers, and judicial police officers scheming to extort drivers. Second, the analysis of an experiment suggests that the scheme subverts service. Third, the scheme appears to be a rational response to the context but its logic is widespread. The findings suggest that coalitions of officials, while being socially costly, can yield large illicit revenue, nuancing the notion of state weakness.

Funding Sources: International Growth Center; Private Enterprise Development in Low-Income Countries

Scope of Project

Subject Terms: State; corruption; weak state

JEL Classification:
      D23 Organizational Behavior; Transaction Costs; Property Rights
      D73 Bureaucracy; Administrative Processes in Public Organizations; Corruption
      O10 Economic Development: General

Geographic Coverage: Democratic Republic of the Congo

Time Period(s): 6/1/2015 – 7/31/2015 (Summer 2015)

Collection Date(s): 6/19/2015 – 7/24/2015 (Summer 2015)

Universe: Interactions between traffic police agents and drivers in the street
Interactions between drivers and judicial police officers in the stations
Interactions between drivers and fine collection agents in the stations

Data Type(s): administrative records data; event/transaction data; observational data; program source code

Methodology

Response Rate: We designed five data collection branches inside the eight stations and 15 intersections (with 63 unique agents) of the target experiment sample to cover the 30 days of the target experiment sample and Day Zero. As described in the next paragraph, the target experiment sample was afflicted by festivities, attrition, and replacements; those affected the data collection activities in this sample, too. Furthermore, on Day Zero, some branches could only cover 14 teams. Thus, the branches were implemented in the 337 team-days of the experiment sample and the 14 teams on Day Zero. This produced an imbalanced panel of 337+14=351 team-days (henceforth, data collection sample), composed of eight stations and their eight managers, 16 JPOs, 16 FCAs, and 63 agents (and 12 replacements) in 15 intersections (and three replacements), shown in Table A2, Column 1 in the article.

In the target experiment sample, closure due to festivities and the removal of two teams for precaution led to the loss of 90 and 38 team-days, respectively (44 and 17 of which were assigned to the encouragement, respectively), resulting in a sample of 337 team-days (the \textit{experiment sample}). First, there were five Sundays and one festivity, leading to the loss of 6 days for each of the 15 non-dropped teams---i.e., 90 team-days. Second, in each of days 3, 5 and 6, we replaced one team, leading to no loss; the teams had been reshuffled; on day 8, we dropped two teams, leading to the loss of 38 (non-festivity) team-days; we had learned that the agents were tense.

Sampling:

The experiment was designed in 2015 in one of the two battalions of Kinshasa's commissariat, covering 7 of its 14 stations as well as Kinshasa's central station. This thus covered the staff of 8 police stations. Such staff was composed of 8 managers, 16 JPOs, 16 FCAs, and 15 teams of agents permanently posted to 15 distinct intersections, totaling 63 agents (13 teams had 4 agents, one of 5, and one of 6).

In the stations, we sampled the universe of interactions between the JPO and the driver (in Branch 1 and Branch 2) as well as between the FCA and the driver (Branch 2). In the street, in each of the 15 intersections, we sampled the universe of interactions between two randomly selected agents and the drivers; in this case, the agents were independently randomly selected each day from the pool of assigned agents for that intersection (which, for each intersection is typically 4).

Data Source:

Data collected as described in article "The Real State: Inside the Congo's Traffic Police Agency." Sources: Branches 1-3.

Collection Mode(s): coded on-site observation; self-enumerated questionnaire

Scales: n/a

Weights:

Branch 1: n/a
Branch 2: n/a
Branch 3: aggregation of events observed from sample of two agents at each intersection accounts for sampling procedure of agents among the pool of (typically) four agents per intersection. A full description of those weights, when they are used, is provided in Supplemental Appendix I of "The Real State: Inside the Congo's Traffic Police Agency"

Unit(s) of Observation: intersection/day traffic properties, team/day arrangements with manager, driver-police officer interaction events

Geographic Unit: intersection

Related Publications

Download this project

Published Versions

V1 [2024-10-23]

Export Metadata

Dublin Core

DDI 2.5

Report a Problem

Found a serious problem with the data, such as disclosure risk or copyrighted content? Let us know.

This material is distributed exactly as it arrived from the data depositor. ICPSR has not checked or processed this material. Users should consult the investigator(s) if further information is desired.