# I. Introduction ecently Peiffer and Rose (2014) asked the question: "Why do some Africans pay bribes and others do not?" This paper attempts to take a closer, more focusedlook at that question by attempting to identify the factors that predict individual level payment of bribes to the police,the institution Peiffer and Rose labeled as the most corrupt in their sample of African countries.The research setting is Kenya, which topped the list ofmost corrupt countries in their African sample. Besides payment of bribes to the police, the survey instrument asked a series of questions regarding payment of bribes for basic public services, healthcare, receipt of documents, sanitation services and school services, basic intra-structure services, hence, intrastructure in the paper's title. The study is unique because it compares those respondents who selfreported paying bribes to the police from those that had the opportunity, but did not report paying a bribe. Scope of Corruption: Corruption can occur on different scales. There is corruption that occurs as small favors between a small number of people (petty corruption). There is corruption that affects the government on a large scale (grand corruption), and corruption that is so prevalent that it is part of the everyday structure of society(systemic corruption). Petty corruption occurs at a smaller scale and within established social frameworks and governing norms. Examples include the exchange of small improper gifts or use of personal connections to obtain favors. This form of corruption is particularly common in developing countries and where public servants are significantly underpaid. This paper will focus specifically on petty corruption, which in Kenya's case meets the criteria to be called systematic. Corruption in Africa: As Chitakunye et al. (2015) indicated, corruption has been described as an intractable global problem from which no nation or region can claim any exemption. Corruption in Africa has been described as a culture (Hope, 2014) and is seen as part of the social fabric of African countries. (Blundo and de Sardan, (2006)Daily life in Africa is governed by the 'petty' corruption of public officials in services such as health, transport, or the judicial system. There is a growing body of literature about the extent and impact of corruption on other societal domains in developing countries, like infrastructure (Kenney, 2006), provision of health care (Young, et al, 2014), foreign aid (Asongu,2012), the public's confidence in public institutions (Clausen,Kraay and Nyiri, 2011). Anoruo and Braha (2005) reported that corruptiondirectly retards economic growth by lowering productivity, and indirectlyby restricting investment. Justensen and Bjornskov (2012)describe corruption as a major source of slow development in Africa, and indicated that corruption and bribery are directly related to poverty because poor people rely on services provided by the government and therefore more likely to be victimsof corrupt behavior by street level government bureaucrats. Which means poor people are those most likely to pay more than their fair share of bribes. While some see corruption as the sand in the wheels of growth, (Wei, 2000), others see corruption as having a positive effect on countries, acting as the grease that moves the economic development process along (Meon and Seekat, 2005) Some have argued that corruption may serve a useful function when it causes commerce to work more effectively. Some scholars take the position that bribery in general may be ethical in cases where it may be a helping hand, but unethical when it is paid to a grabbing hand where nothing is offered in return for the payment (Colombatto, 2003;Egger and Winner, 2005;Houston, 2007). Wong and Beckman (1992) developed a point system to determine whether the helping or grabbing hand predominates. Corruption and infrastructure in Kenya: The rationale for this paper is provided by several earlier studies generated by the Afrobarometer Project, the data source used in this research. Besides Peiffer and Rose (201), these include Justensen and Bjornskov (2012), Richmond and Alpin (2013)and Bleck and Michelitch (2015). Peiffer and Rose indicated that the Afrobarometer studies show a range of levels of bribery, with Kenya showing the highest percentage of respondents paying a bribe, 49 percent. Nigeria was tied with Uganda second place on that list with both reporting 42 percent of respondents paid a bribe.They also indicated that the police were identified as the most corrupt institution in Kenya, a finding echoed over and over in the African literature. Peiffer and Rose stressed that one major value of surveys such as Afrobarometer is that they focus on the experience of a representative sample of citizens in the encounters they have with lowlevel officials delivering public services in their local areas. One recent interest has been infrastructure, or the lack there of in Africa. Afrobarometer has published several recent papers on this topic One important paper was produced by Leo, Ramachanddran and Morello (2015). They noted that the need for infrastructure improvements is a top level economic, political, and social issue in nearly every African country. They also note there is an extensive academic and policy literature about the impact of infrastructure deficits on economic and social indicators. Yet, very few studies have examined citizen demands for infrastructure. Leo, Ramachandran and Moeello produced infrastructure data from 33 countries included in Afrobarometer Round 5. The purpose was to move toward a basic understanding of service availability as perceived by citizensin each country. They found a predictable pattern of infrastructure services across income levelslower-income countries have fewer services. The survey data also allowed the authors to observe the sequencing of infrastructure services. While survey respondents were most concerned with jobs and income-related issues, they were also concerned with the availability of infrastructure, specifically transportation and sanitation. These are priorities which transcend demographic factors, including gender and location (urban/rural). Another recent Afrobarometer paper by Bleck and Michelitch (2015) dealt with a single country,Mali, and thus serves as a precursor to the purposes of this paper. That paper provides the perspectives of rural Malians living on the border of state-and rebelcontrolled territory during a major crisis in the country, a coupe and a secessionist insurgency. Using a mixed methods approach, Bleck and Michelitch found it was not the political situation that rural villagers found.as "the crisis". Rather it was the unmet needs for public services and infrastructure. The state breakdown mattered less because the state had not been present in the first place. Rather than the state, villagers were largely reliant on local traditional authorities. The authors concluded that the salience of villagers' concerns about public services and infrastructure, as well as general insecurity related to basic needs, were consistent with the rest of Africa. In fact their findings echoed cross-national Afrobarometer data (2012-2013) on public service provision across rural and urban areas. Their conclusion was that for all sub-Saharan African countries, the urban-rural gap is large, with absolute levels of rural provision low, and countries are inconsistent in provision of services across all infrastructure indicators. # II. Method and Materials The study's data source is Afrobarometer, an independent, non-partisan research project that conducts surveys in more than 30 African countries, and they are repeated on a regular cycle. This study is based on Round 5 which the Project conducted in Kenya.in 2011.Like the previous surveys, round 5 consisted of face to-face interviews with Kenyans 18 years of age and older. The sample consists of 2,399 respondents based on interviews conducted in multiple languages. The sampling frame included all of Kenya's provinces and the final sample provides estimates of the national population of all adults in Kenya that is accurate to within a margin of error of plus or minus 2 percentage points at a confidence level of 95 percent.The sampling procedures that are used in all Afrobarometer surveys are explained in detail in Bratton, Mattes and Gyimah-Boadi (2005). # Measures and Statistics: The dependent variable: The study's dependent variable is payment of a bribe to the police. Survey respondents were asked a series of questions about payment of bribes. The questions were asked as "in the past year, how often, if ever, have you had to pay a bribe, give a gift or do a favor to government officials?" Fixed responses for this and the other questions in the series were as follows ; never, once or twice, a few times, often, no experience with this the last year and don't know. Responses to the police question were coded as follows; never and no experience with this in the last year were coded 0 (no), and once or twice, a few times and often became 1 (yes). This dichotomous variable provides the basis for the logical regression analysis presented below. The independent variables: The study's independent variables include the other types of services on the list where possible bribes could have been paid.These included to obtain a document, receive water or sanitation services (sewer), receive treatment in a hospital or health clinic, or to receive admission or extra services in schools for a child. All of these measures were coded in the same manner as the study's dependent variable, either a yes (1) or a no (0).The interviewer and supervisor attempted to verify the availability of those public services in the respondent's local area. Respondents were also asked to rate whether obtaining certain public services was easy or difficult. They were also asked about their perceptions of the police, did the respondent trustthem?Do you think the police are corrupt? Other questions related to the police were recorded by the interviewer and verified by the supervisor, namely whether the police were visible in the area and whether a police station was located in the area. Respondents were asked whether they had been a crime victim within the last year, either a property crime victim (was something stolen from their house?) or were youor someone in your family a violent crime victim (Were you or someone in your family physicaslly attacked in your home?These Kenyan respondents were not asked to report their income in the Afrobarometer survey.As Bratton (2008) indicated this is because many citizens in poor countries operate in informal marketswhere cash transactions, including income, are unrecorded and difficult to measure.Instead, this research used what is called an Asset-based WealthIndex, a summed index created from three questions that ask abouthousehold assets. The survey asks respondents: "Which of these things do you personally own: A radio?A television? A motor vehicle, car or motorcycle?" Responses to these questions are binary (0=don't own; 1=own), Some of the study's control variableswere measured by a single item, like age, gender, religion, employment status, residence (rural or urban),and education, which was collapsed into 5 categories, which ranged from a none category to college/university graduation and graduate school.The samples basic demographic characteristics are displayed in Table 1. Table 1 showed that this Kenyan sample was rather young, with only 16 percent over 50 years of age, with43 percent found in the between 30 and 49 years of age group. Afrobarometer samples are drawn to equally reflect the sexes, and Table 1 shows that was achieved in this Kenyan sample. The sample was primarily Christian, 86 percent, rural, 62 percent and over half the sample was unemployed; 24 percent reported having full time employment. There was a range of responses to the educational attainment question, from none or informal education only. 7 percent to college or graduate school completion, 5 percent. Some high Year 2015 # ( H ) school or completion of high school accounted for40 percent of the sample. A higher percentage of respondents reported owning none of the possessions listed in the question about ownership, 17 percent, compared to 9.4 percent who reported owning a radio ,TV and a motor vehicle (including motorcycles). Note that race was not included in Table 1, because there was no variation by race in this sample, with 98.9 percent of the respondents listed as Black Africans. The next step was to examine the payment of bribes measures and the questions which revealed whether certain infrastructure services were available in the local area. Those indicators are displayed in Table 2. Table 2 shows that 32.1 percent of this Kenyan sample reported paying a bribe to the police within the last year. Another 38.9 percent reported paying a bribe to obtain a document, 17.5 percent reported bribes for sanitation services, 27.8 for health care, and 17.2 percent for school admission or extra assistance for a child at school. In terms of the availability of infrastructure in the local area, 71.3 percent of respondents lived where the electricity grid was available, 45.3 percent lived where piped water was available and 15.3 percent resided in an area which had a sewer system. The next series of questions were those that asked respondent perceptions about either the difficultyor ease in accessing or receiving public and police services, whether there was a police station in their local area, whether police were visible in the area, and finally whether in the last year they had been victims of either a property or violent crime in their homes. The responses to these questions are displayed in Table 3. Table 3 shows that, overall these Kenyan respondents felt receiving public services was difficult. In terms of obtaining documents, 80.7 percent thought it was difficult, 70.9 percent felt the same about acquiring household services, like water and electricity, 74.3 felt it wasdifficult to place children in public schooland 58.0 percent thought receipt of medical services was difficult.In terms of the police,74.3 percent thought receiving help from the police was difficult. Also included in Table 3 were indicators of police visibility and proximity. Police were visible in areas where 32.0 percent of the respondents resided and there was a police station in the local area of 36.3 percent of the respondents. The questionnaire did not include any measure of direct involvement or contact with the police, but respondents were asked about crime victimization within the last year. About one third of the respondents, 33.0 percent, had something stolenfrom their homes and 11.1 percent or family members had been victims of violent crimes in their home within the last year. The final task in the analysis was to conduct a logistical regression analysis. All of the variables included in Tables 1 through 3 were entered into the regression equation. The results are displayed in Table 4. Table 4 reveals that nine variables reached statistical significance in the logistical regression analysis; the analysis produced a pseudo R2 of .22. The most import finding in Table 4 was that all four of the payment of bribes indicators were included in the statistically significant group. The strongest predictor was payment of bribes to obtain documents,(Z=9.59), followed by payment of bribes for school services. (z=6.34). Payment of bribes to receive healthcare services and sanitation services were also significant, but not as strong (Z=2.55 and Z=2.51 respectively). As suggested by the literature, gender was a strong predictor of the payment of bribes to the police, (Z= -5.86) : men pay bribes. Trust of the police was also significant. There were several surprises. One was the fact that difficulty receiving sanitation services was significant, and none of the other questions in that series about ease or difficulty of obtaining services were significant. Perhaps the most unexpected finding in Table 4 was that being the victim of a violent crime was significant.( Z=4.62). This question and the one about being a property crime victim, which was not significant, were included in the logistical regression analysis because both measures suggest direct police contact # Volume XV Issue IV Version # III. Discussion This paper found support for the belief that a true culture of corruption exists in Kenya and that was an important finding. More than that, the study used data generated at the individual respondent level, which isunique. Studies concerned with corruption usually rely heavily on Transparency International and its Corruption Perceptions Index (CPI) as a major data source (available at http://.www.transparency.org). There were some surprises in this paper, and the issue of respondents self-reporting crimes was one of these; Kenyan respondents did report payment of bribes to a greater extent that might have been expected. Perhaps this reporting of payment of bribes should be considered surprising, but the volume of those responses suggests,as far as Kenya is concerned, the question asked at the beginning of the paperperhaps shouldbe reframed as"why doesn't everyone in Kenya pay bribes.?" This paper clearly suggests that most Kenyans do pay bribes. One interesting finding was the lack of infrastructure in Kenya, especially sewer and water. Note The Factors that Predict Systematic Police Bribery in Kenya was possible. The issue of violent crime victimization will be addressed below. that only about 15 percent of the respondents lived where a sewer system was available. And the fact that difficulty in obtaining household services was a Year 2015 # ( H ) The Factors that Predict Systematic Police Bribery in Kenya significant predictor of payment of bribes to the police and others was not an unexpected finding. The finding that requires more attention is the fact that violent crime victimization was a predictor of payment of bribes to the police. There are several ways to approach explaining that finding. One of them is to suggest the need for an improved measure regarding paying the police a bribe. As presently presented the question asks if payment was to police passing a point, avoiding a fine or arrest. The police in Kenya and other African countries are known to routinely ask motorists for bribes to pass checkpoints and/or road blocks. That is not the same as avoiding a fine or arrest, especially an arrest. So a better question about avoiding an arrest in warranted and would help clarify why the bribe is paid to the police. If respondents seem willing to report payment of bribes, why not ask the reason the bribe was paid? A second explanation to the payment of bribes to police is directly related to police corruption. When payment of police bribes was cross-tabulated with being a violent crime victim, it was discovered that more than half of the violent crime victims had paid the police a bribe. (130 of 258 or 50.4 percent). Although not as strong, property crime victims also paid a higher percentage of bribes to the police (288 of 778 or 37 percent), higher than would be expected. Two explanations follow from those findings. On the one hand, it suggests that if you get in an altercation, you may be the victim and not the perpetrator if you pay the police a bribe or more money than the other person. If you are a property crime victim, it may be that you think you need to pay a bribe if you expect the police to look for or even return your stolen property. None of these scenarios casts the police in a good light. # IV. Conclusion The findings related to the payment of police bribes and payment of bribes in other service provision areas were exceedingly strong, and support the notion that Kenya has a bribery culture. However, what is worrisome is that these findings do not apply to the entire country, and are limited to areas of Kenya which have more developed infrastructures. The lack of infrastructure development in Kenya suggests that as infrastructure is more thoroughly developed and implemented, the bribery culture will only grow in Kenya in the future, and everything possible must be done to institute anti-bribery programs. Starting to address the street level interaction between low level bureaucrats (public service providers) appears to be the most promising to begin to address the corruption problem. Police are only part and parcel of a much larger culture of corruption problem in Kenya. 4![I Global Journal of Human Social Science © 2015 Global Journals Inc. (US)The Factors that Predict Systematic Police Bribery in KenyaTable Logistic regression with Violence Victimization as the Dependent Variable Number of observations = 1 868 Chi square = 576.85 Probability = .000 Pseudo R2 = .22](image-2.png "4 :") 1Variable 2Variable 3Variable © 2015 Global Journals Inc. (US) - ## Global Journals Inc. (US) Guidelines Handbook 2015 www.GlobalJournals.org * Corruption and economic growth: the African experience EAnoruo HBraha Journal of Sustainable Development in Africa 7 1 2005 * On the effect of foreign aid on corruption SAAsongu WP/12/031 2012 * On the primacy of weak public service provision in rural Africa: Malians redefine 'state breakdown' amidst 2012 political crisis -Afrobarometer Working Paper 155 JaimieBleck KristinMichelitch 2015 * MichaelBratton RobertMattes &EGyimah Democracy, and Market Reform in Africa 2005 Cambridge University Press * Poor People and Democratic Citizenship in Africa MichaelBratton Poverty, Participation and Democracy AnirudhKrishna New York Cambridge University Press 2008 * Everyday Corruption and the State GBlundo JPOlivier De Sardan Citizens and Public Officials in Africa N. B. Arifari and M. Tidjani Alou 2006 Zed Books * Transnational Analysis of the Impact of Corruption on Development in Africa: A Review of Literature PChitakunye¹ DJOjochenemi EDerera ATarkhar 2015 * Evidence on corruption as an incentive for foreign direct investment PEgger HWinner European journal of political economy 21 4 2005 * Kenya's corruption problem: causes and consequences KHope Commonwealth and Comparative Politics 52 4 2014 * Can corruption ever improve and economy DAHouston Justesen KMogens ChristianBjørnskov Exploiting the Poor: Bureaucratic Corruption and Poverty in Africa -Afrobarometer Working paperno 2007. 2012 27 * Measuring and reducing the impact of corruption in infrastructure CKenny 2006 World Bank Policy Research Working Paper * BenjaminLeo RobertMorello VijayaRamachandran The face of African infrastructure: Service availability and citizens' demands -Afrobarometer Working Paper 2015 * RMattes MBratton YDavids Poverty Survival, and Democracy in Southern Africa 23 2003 Afrobarometer Working Paper * Does corruption grease or sand the wheels of growth? PGMéon KSekkat Public choice 122 1-2 2005 * Why do some Africans pay bribes while other Africans don CarynPeiffer RichardRose 2014 Afrobarometer Working Paper * Governments falter in fight to curb corruption: the people give most a failing grade SRichmond Alpin C Transparency International 2013. 2015 Afrobarometer Policy paper No 4.oAst a failing 18 * How taxing is corruption on international investors SJWei Review of economics and statistics 2000 82 * An applied ethical analysis system in business AWong EBeckman Journal of Business Ethics 11 3 1992