Protest and property crime: political use of police resources and the deterrence of crime

The tradeoff in the allocation of police resources between drug enforcement and non-drug crime is not unique to the United States. See, for instance, Mendes (2000) and Miron (2001).

Some research, however, has maintained that drug users are no more likely to commit non-drug-related crimes than are non-drug users (e.g., Benson et al. 1992).

Note that political protests could also be endogenous to Index I crimes. In the United States, for instance, political protests often are accompanied by theft, arson, and even vandalism. In Korean culture, however, those kinds of behaviors are not significant because people engaging in political protests generally are not members of crime-prone groups. If political protests are indeed positively associated with Index I crimes in South Korea, then the tradeoff effect may be even larger than the results of our study indicate.

As an anecdote, the Seoul Metropolitan Police Agency temporarily reassigned police forces from public security (e.g., protest control) to public safety (e.g., patrolling and criminal investigations) between December 2014 and January 2015. As a consequence, the incidences of five types of serious crimes (homicide, robbery, rape and sexual harassment, theft and violence) fell by 34.4% compared to the same period the previous year. In particular, the reduction in property crimes, such as larceny-theft and burglary, was substantial because, for instance, police patrols in residential areas were reinforced. That anecdote suggests that even the temporary reallocation of police forces seems to deter crime, potentially because police forces are effective across different functions (e.g., across public security and public safety).

Similarly, prosecutors pursue their (bureaucratic) self-interests in prosecuting cases. For instance, they may prefer cases that either require less time to prepare for trial or benefit their future career prospects (e.g., Boylan 2005; Gordon and Huber 2009).

The policies range from the decision to send troops to Iraq (2004) and the importation of US beef (2008) to the Miryang transmission tower projects (2014). Other major incidences include one on the Yangju highway (2002), in which a US armored military vehicle struck and killed two teenage schoolgirls; the impeachment of President Roh Moo-hyun (2004); the Yongsan incident (2009); meddling in the presidential election by the National Intelligence Service (2012); the Seoul City Employee Spy case (2013); the Sewol ferry disaster (2014); and the impeachment of President Park Geun-hye (2016–2017).

The Korea National Police Agency (central agency) and the 17 local agencies are subordinate to the Minister of the Interior.

This sometimes results in disproportionately large police forces mobilized for certain protests. For instance, Arnold Fang of Amnesty International said, “The force used by police at the Miryang protest was disproportionate and in breach of international standards.” Note that the Miryang protest (2014) took place in the second year of President Park’s administration.

The share of police officers in security control relative to police officers in crime investigation increased from 45.2% in 2001 to 54.1% in 2015. During the same period, the probability of arrest for conventional crimes (including property and violent crimes) fell from 78.2 to 74.9%. The reduction in arrest rates may explain why conventional crimes rose at an annual rate of 1.9% during the same period. These data were obtained from the Police Statistical Yearbook, published by the Korean National Police Agency, and the Annual Crime Reports, published by the Supreme Prosecutors’ Office of Korea.

All crime data and deterrence variables were obtained from the Annual Crime Reports. Demographic and socioeconomic variables were obtained from Statistics Korea (http://​kosis.​kr).

Other categories of property crimes include fraud, embezzlement and vandalism. See Kim and Kim (2015) for detailed accounts of these subcategories.

All data on protests and police deployment were obtained from the Police Statistical Yearbook by the National Police Agency. Some of the missing values were collected through the Open Information System (www.​open.​co.​kr) and interviews with local police agencies. The Korean National Police Agency classifies six types of police mobilization for security purposes: protests, guard/security, emergency guard, congestion, disasters and elections. We focus on protests and guard/security mobilizations because those two types of deployment are most likely to be politically relevant in Korea. For instance, we do not include police security deployments for elections because major elections take place only every two to four years, and elections in Korea usually are peaceful events that require limited police presences.

A local police agency consists of various departments working in cooperation, including officers assigned to the public security, criminal and special investigation, national security and narcotics units.

Although the decision to reallocate police resources is made at the local level, the central agency has some influence on the actions of local police agencies. That influence opens up the possibility that the central agency may allocate more police resources to areas where political opposition is greater if protests are frequent. Another possibility is that more police resources could be allocated to locations where the regime is supported strongly (to strengthen that support). If so, the tradeoff effect will not exclusively be local because the other areas from which police are drawn should have less crime control regardless of political protests in those areas. However, reallocation of police resources across cities and regions is not common.

President Park Geun-hye was impeached on March 10, 2017.

We first collected the number of prosecutors from all 58 district prosecutors’ offices. We then grouped them into 13 metropolitan areas to calculate the probability of prosecution.

Since the early 1970s, many studies have found a positive effect on crimes of certain age groups in a predominantly male population (Freeman 1996).

The recent literature has focused on unemployment rates among certain crime-prone groups, such as young people, because the aggregate unemployment rate may not identify the marginal criminal (Raphael and Winter-Ebmer 2001; Gould et al. 2002; Lin 2008; Mustard 2010).

Previous studies have emphasized that alcohol consumption is an important predictor of violent crimes (e.g., Saridakis 2004).

We thank an anonymous reviewer for suggesting the IV approach.

We do not include time dummies because of multicollinearity with EARLY _t . We also included both linear and quadratic time trends (not reported). The results did not affect our main results.

The impact of PROTEST on CR is potentially endogenous, however. For instance, a lack of crime control (i.e., higher CR) and more political protests may both reflect social instability. Another possibility is that if political protests are about the government’s handling of crime, a potential endogeneity problem could arise. However, most political protests arose over government policies involving sending troops overseas, international trade, infrastructure projects and impeachment of the president. Very few, if any, protests were about the handling of crime during our sample period. In any case, excluding PROTEST1 and PROTEST2 did not change the main results (not reported).

For brevity, we only report the results for POLIT1. Using POLIT2 does not change the main results (available upon request to the authors).