Introduction: Country Classification
Individual firms’ sales losses can be determined by a variety of factors. These factors can be divided into two general categories: internal determinants and external determinants (Alifano, Corradi, and Distaso, 2019). While internal factors are determined using a detailed analysis of firms, external factors can be identified using comparative analysis between different groups of firms in various political, economic, and social settings. The underlying advantage of such comparative analysis is that researchers do not need to collect primary data, as secondary data is available for public use.
The present paper aims to conduct a comparative analysis of factors affecting sales losses associated with theft and vandalism. The primary dataset used as the source of secondary data is “Losses due to theft and vandalism (% of annual sales of affected firms)” provided by the World Bank (2021a). The paper also uses supplementary datasets provided by World Bank to identify the factors affecting sales losses.
The original dataset was in two depending on the country’s group. All the countries were classified into “developing” and “developed” based on Human Development Index (HDI). According to Bacchus and Manak (2020), a country can declare itself developed if its economy is highly developed and technologically advanced. Additionally, the country’s HDI should be relatively high, unemployment should be low, and conditions of living should be improved (Paprotny, 2020). It is usually considered that an HDI of 0.8 and above is enough for a country to be classified as developed (United Nations Development Programme, 2020).
The comparative analysis of factors affecting the sales losses due to theft and vandalism is expected to determine if there are differences in the behaviors of populations of developed and developing countries. In particular, the analysis will help to understand if firms in developing countries are at higher risk of revenue losses from theft and vandalism and if factors affecting the matter differ. The results of the analysis will contribute to a deeper understanding of the differences between developing and developed countries.
Literature Review
The determinants of property crime are widely discussed by scholars and policymakers. For instance, Rosenfeld and Messner (2009) conducted a comparative analysis of factors affecting the decline in crime rates in the US and several European countries. The researchers hypothesized that economic determinants, such as unemployment, consumer confidence, GDP per capita, and imprisonment levels, affected the prevalence of property crime in all the countries under analysis. All the data was acquired from public sources, such as Eurostat and the Organization for Economic Cooperation and Development. The authors utilized Pearson’s R and regression analysis to identify the effects of the factors on crime levels. The results revealed that the determinants of burglary were similar in European countries and the US. In particular, crime rates were found to be affected by economic changes and increased imprisonment rates.
Crawford (2013) compared different groups of countries in a narrative review of reasons for affecting the prevalence of crime. The primary claim of the author was that effective crime prevention policies affect crime rates in different countries. The researcher utilized data from publicly-available sources on crime rates and policy descriptions to compare third-world countries to European countries. The result revealed that effective crime prevention practices were one of the crucial determinants of crime rates in all countries (Crawford, 2013). The effectiveness of practices was associated with the level of economic development of different countries. It should be noticed that Crawford (2013) utilized qualitative methods to answer the primary research question. Therefore, the reliability of the study’s results is questionable.
Ghani (2017) conducted a comparative study of crime rates between Malaysia and Nigeria. The primary hypothesis was that urbanization levels were associated with higher crime rates in both countries, even though Malaysia was developed and Nigeria was developing. The researcher utilized publicly available databases to collect relevant quantitative data. However, only descriptive statistics were used to analyze the data. The results revealed that urbanization measured in the density of the population was positively associated with the growth of both property and violent crimes (Ghani, 2017). Additionally, Ghani (2017) argued that unemployment, poverty, poor political governance, and weakness of law-enforcement bodies were determinants of a high prevalence of crime. In summary, the majority of studies agree that economic factors and law enforcement policies are central determinants of the prevalence of property crimes.
The Determinants of the Regressand
The literature review revealed that there is a wide variety of aspects that can be considered determinants of the prevalence of property crime. In particular, crime prevalence can be explained by economic factors (poverty rates, GDP per capita, and unemployment), social factors (urbanization), and policies (strictness of punitive measures) that were correlated with property crime prevalence. The present paper assumes that an increased number of property crimes is associated with high percentages of firms that experienced significant revenue losses due to crime and vandalism. Based on the literature review and this assumption, the following variables were identified.
Regressand: Average Percentage of Losses to Theft and Vandalism
The dependent variable was identified as the average percentage of losses to theft and vandalism. The dataset provided by World Bank (2021a) included data on the percentage of losses due to theft and vandalism between 2006 and 2020. However, the data was often missing for individual years, which would make the analysis inconsistent. Therefore, it was decided to use average percentages to deal with the problem of data omission. The countries with no information on the regressand were excluded from the analysis.
Determinant 1: Average GDP per Capita
The first determinant of the regressand was the average GDP per Capita. In order to be consistent with the regressand, the average value for years between 2006 and 2020 was used from the dataset provided by World Bank (2021b). Rosenfeld and Messner (2009) utilized GDP per capita as the primary determinant of the economic development of a country. Economic growth can explain up to 50% of the decline in property crimes (Rosenfeld and Messner, 2009). Thus, the following hypotheses were identified:
- Hypothesis 1A: Average percentage of losses to theft and vandalism is negatively correlated with economic growth, described as GDP per capita in developed countries.
- Hypothesis 1B: Average percentage of losses to theft and vandalism is negatively correlated with economic growth, described as GDP per capita in developing countries.
- Hypothesis 1C: Average percentage of losses to theft and vandalism is negatively correlated with economic growth, described as GDP per capita in all countries regardless of their classification.
Determinant 2: Average Unemployment Rate
The second determinant of the regressand was the average unemployment rate. Similar to the first determinant, an average unemployment rate between 2006 and 2020 was taken from the dataset provided by World Bank (2021c) to be consistent with the regressand. According to Ghani (2017), unemployed people are more likely to commit property crimes to meet their essential everyday needs in food, clothing, and shelter. Additionally, unemployment often means that people have nothing to do, which can lead to the organization of gangs engaging in property crimes (Ghani, 2017). Thus, the following hypotheses were formulated:
- Hypothesis 2A: Average percentage of losses to theft and vandalism is positively correlated with the average unemployment rate in developed countries.
- Hypothesis 2B: Average percentage of losses to theft and vandalism is positively correlated with the average unemployment rate in developing countries.
- Hypothesis 2C: Average percentage of losses to theft and vandalism is positively correlated with the average unemployment rate in all countries, regardless of their classification.
Determinant 3: Average Urban Population
The final determinant utilized in the present paper was urbanization, measured as the average urban population. Similar to the two previous independent variables, average urban populations for the past 15 years were taken from the dataset provided by World Bank (2021d) to be consistent with the dependent variable. According to Ghani (2017), the number of property crimes is higher in urban areas in comparison with rural zones. This can be explained by the ability to form gangs and engage in organized crime activities. Thus, it is expected that countries with higher urban populations would have higher percentages of sales losses due to property crimes. The following hypotheses were identified based on this assumption:
- Hypothesis 3A: Average percentage of losses to theft and vandalism is positively correlated with the average urban population in developed countries.
- Hypothesis 3B: Average percentage of losses to theft and vandalism is positively correlated with the average urban population in developing countries.
- Hypothesis 3C: Average percentage of losses to theft and vandalism is positively correlated with the average urban population in all countries, regardless of their classification.
Regression Results and Interpretations
Analysis Description
Regression analysis was conducted to identify what factors affected the percentage of losses to theft and vandalism. In total, three models were assessed using three different datasets. The first dataset included only developing countries (n = 111), the second dataset included only developed countries (n = 27), and the final dataset included all the countries. All the countries with omissions in data were excluded from the final dataset; thus, the total number of observations decreased from 268 to 138, which implies that only 51.5% of countries were analyzed. The majority of countries were excluded due to the lack of values for the regressand.
In total, the regression analysis included the analysis of the dependent variable against three explanatory variables and a dummy variable, which indicated if the country was developed or not according to the HDI index. The model for the regression analysis was as follows:
Where y = Percentage of Losses to Theft and Vandalism
- x1= GDP per Capita
- x2= Unemployment rate
- x3= Urban Population (in thousands)
- x4= Dummy variable for developing/developed
The results of the regression analysis are demonstrated in Table 1 below.
Table 1: Regression Analysis Results
Notes: *, **, and *** indicate significance at the 10%, 5%, and 1%, respectively. The numbers in parentheses are p values.
Discussion of the Results
The majority of the results were inconsistent with the literature review, as the majority of the hypotheses were rejected. First, the unemployment rate in developing (p = 0.274), developed (p = 0.449), and all counties (p = 0.602) had no significant impact on the percentage of firms affected by sales loss due to theft and vandalism, which implies that hypotheses 2A, 2B, and 2C should be rejected. Such results are inconsistent with the literature review, as Rosenfeld and Messner (2009), along with Ghani (2017), stated that unemployment was a significant determinant of property crimes.
Second, the urbanization level seemed to have no effect on the regressand in developing (p = 0.213), developed (p = 0.415), and all analyzed countries (p = 0.15). This implies that hypotheses 3A, 3B, and 3C should be rejected. The findings are inconsistent with Ghani (2017), as no significant correlations between urbanization and the level of property crimes were found. Such inconsistency can be explained by several reasons. First, the measure of urbanization may have been inappropriate, as real values were compared to the percentages. Better results may have been obtained if urbanization was measured as a percentage of the population. Second, the suggestions concerning the correlation between urbanization and crime prevalence made by Ghani (2017) were based on experts’ opinions rather than on robust statistical analysis. Thus, statements made by Ghani (2017) may be biased.
Third, the data analysis found only partial support for the notion that GDP per Capita affects the level of property crime. In particular, the correlations between the two variables were found significant for developing (p < 0.001) and all countries (p = 0.03), while there were no significant correlations between the regressand and GDP per Capita and percentage of losses to theft and vandalism (p = 0.591). Thus, hypotheses 1B and 1C were accepted, while hypothesis 1A was rejected. Such results were inconsistent with Rosenfeld and Messner (2009), as the research was based on a comparison of developed countries. This can be explained by the fact that once one GDP per capita reaches a certain level, it stops having a significant effect on the regressand.
It should also be noticed that the predictive ability of all the regression models was low, as adjusted R2 varied between 0 and 15.87%. This implies that the models were incomplete, which implies that additional variables should be included in further analysis. It was initially expected to include information on the number of imprisonments and the percentage of the budget spent on law enforcement. However, such data was not found in the open access. Future research should focus on obtaining such information and adding it to the model to improve the results.
In addition to the possible reasons for the inconsistency of results with previous research mentioned above, other explanations should be considered. It is crucial to note that the present research was based on the assumption that an increased number of property crimes is associated with high percentages of firms that experienced significant revenue losses due to crime and vandalism. Since no empirical evidence was found for such an assumption, it may be flawed. Moreover, many countries were excluded from the analysis due to the absence of data, which may have resulted in biased results.
Summary and Policy Discussions
Summary
The present paper aims to conduct a comparative analysis of factors affecting the percentage of firms that experienced losses due to theft and vandalism in developing and developed countries. Nine hypotheses were formulated based on the literature review. After obtaining data from open sources, three models were assessed using regression analysis. The results revealed that GDP per capita had a significant effect on the percentage of firms that experienced losses due to theft and vandalism in developing countries. However, no such correlation was found in developed countries. Moreover, unemployment and the amount of urban population did not have any significant effect on the regressand.
Strengths and Weaknesses
The primary strength of the present paper is the data utilized for the investigation. All the data is taken from a highly reputable source, which implies that the collection methods were adequate. Additionally, the paper utilized robust methods for data analysis, which were used in previous research. Thus, the reliability of methods is high, which is associated with increased reliability of results.
However, several weaknesses should be acknowledged to avoid bias. First of all, it should be mentioned that the operationalization of concepts may have been inadequate, as urbanization was operationalized as urban population instead of the percentage of the urban population. Additionally, only three possible determinants together with a dummy variable were used. This number of determinants seems inadequate for describing such a complicated matter as losses due to theft and vandalism.
Policy Suggestions
The results demonstrate that developing countries can decrease losses associated with property crimes by increasing their GDP per capita. The economic determinant can be improved by investing in innovation and technological advances in production. Thus, the government of developing countries should focus on the promotion of innovative products in the country. As for the developed countries, if the government needs to decrease losses associated with theft and vandalism, they should invest in developing crime prevention mechanisms, as improving GDP per capita, decreasing unemployment, and decreasing urbanization appear inappropriate for this matter.
References
Alifano, D., Corradi, V., and Distaso, W. (2019) The determinants of operational risk losses. SSRN.
Bacchus, J. and Manak, I. (2020) The development dimension: What to do about differential treatment in trade.
Crawford, A. (2013). Crime prevention policies in comparative perspective. Abingdon-on-Thames: Routledge.
Ghani, Z. A. (2017) ‘A comparative study of urban crime between Malaysia and Nigeria’, Journal of Urban Management, 6(1), pp. 19-29.
Paprotny, D. (2021) ‘Convergence between developed and developing countries: A centennial perspective’, Social Indicators Research, 153(1), pp. 193-225.
Rosenfeld, R., & Messner, S. F. (2009) ‘The crime drop in comparative perspective: the impact of the economy and imprisonment on American and European burglary rates’, The British Journal of Sociology, 60(3), pp. 445-471.
United Nations Development Programme (2020) Human development reports.
World Bank (2021a) Losses due to theft and vandalism (% of annual sales of affected firms).
— (2021b) GDP per capita (constant 2010 US$).
— (2021c) Unemployment total.
— (2021d) Urban population.