Размещено на http://www.allbest.ru/

Размещено на http://www.allbest.ru/

The impact of corruption on migration

Егорова Мария Владимировна

ABSTRACT

migration economic corruption

The purpose of this work is to determine the political factors of bilateral migration. The Corruption Perceptions index, the Control of corruption index and the Rule of law index are used in this study in order to check the political effect on migration. The results of the corruption effect were established on the basis of PPML regressions for 1995-2015-year data for bilateral migration with 5-year intervals. We also analyze the gender migration and the migration from countries of different income categories in order to evaluate the differences in the effects for the migration of rich, middle-income and poor people and to check whether the effect of corruption continues to exist when we consider the migration into regions with high GDP per capita. The results show that the corruption in the country of destination is a significant factor for migrants. Migrants tend to move to the less corrupted governments with better political regulations. For the females the effect of corruption in the country of destination is higher than for males. When we consider the migration of different income categories, it comes out that the corruption is out of importance for the migrants from low-&medium- income countries to high-income countries, while it is a significant factor for migrants from high-income to high-income countries. The obtained results are of great relevance for the government authorities that have a purpose to increase the inflow of working power or to decrease the level of brain drain from the home country.

Key words: Bilateral migration; Corruption; Migration outflow; Migration inflow; Gravity model.

CONTENTS

• INTRODUCTION
• CHAPTER 1. THEORETICAL FRAMEWORK
• CHAPTER 2. EMPERICAL RESULTS
• 2.1 Data
• 2.2 Methodology
• 2.3 Results and Discussions
• 2.4 Country-specific effects
• 2.5 Robustness check
• CONCLUSION
• REFERENCES
• APPENDIX

INTRODUCTION

Nowadays the increment of the migration flow continues to increase by each year. According to the United Nations Organization, the number of migrants all over the world in 2000 accounted for 173 millions of people, in 2010 it was 220 millions of people and in 2017 it reached the value of 258 millions of people. These migration movements affect the economy of the countries of origin and the countries of destination due to their impact on the number of working population and GDP per capita in both countries, so the reasons for migration and the methods of its regulation should be analyzed (Ahmad, Arjumand, 2016).

Currently, the migration flows have become a vitally important issue for the heads of the government. The acuteness of the problem has risen due to the European migrant crisis which occurred in the 2015. The significance of this problem can be explained by the fact that the migrants serve as a workforce that either would increase the amount of goods and services in the country of destination or decrease it in the home country (Carling, Paasche, Siege, 2015). The flow of migrants is determined by many different reasons. One of such main factors that affect the migration flow is the political environment which is supposed to be very significant reason for the migrant decision to enter the foreign or leave the home government. It can be explained by the fact that migrants, while making this decision, are limited by politics in terms of freedom of movement- they should pay passport and visa costs and follow the rules of the government which they enter (Jain, 2001). Also, the political regulations in the home countries can make original residents stay and work or attract some new foreign workers which also affects the migrant flow (Metelev, 2014).

As for the results of our study, we analyzed the relationship of the migration and political factors that can have an effect on the migrant's decisions about whether to enter the foreign government or leave the home country. It was revealed that there exists a positive relationship between the Corruption Perceptions index (CPI), Control of corruption index, Rule of law index and the migration flow. With the help of PPML regressions which allowed us to avoid heteroskedasticity we found out that there is a significant positive effect of CPI, Control of corruption and Rule of law in the country of destination on the number of migrants between the country of origin and the country of destination. So, the anticorruption regulations in the country of destination serve as a pull factor for migrants. In addition, it was revealed that for females the effect of these political factors is bigger than for males. In order to check whether the GDP per capita is a more important factor for migrants that can drive out the effect of political indexes in the country of destination we considered the migration from all countries to the high-income countries and then divided this migration on the migration from low-&medium-income countries to high-income countries and the migration from high-income to high-income countries. It came out that the effect of CPI, Control of Corruption and Rule of law in the country of destination disappeared for the migrants from poor and medium-income governments that enter the foreign government in order to find a job or get more money. As for the migrants from high to high-income countries, the CPI, Control of Corruption and Rule of law in the country of destination are important factors for them, so they care about the corruption environment while migration to the foreign government.

CHAPTER 1. THEORETICAL FRAMEWORK

The migration flows have become an integral part of the world today, and their growth rates only increase by each year. That is why migration is a widely discussed process. Nowadays, there is a limited amount of literature that is dedicated to political determinants of the gender and regional bilateral migration from different income regions. At the same time, there is a big number of articles connected with the theme of worldwide migration and its political factors. So, some articles dedicated to the migration determinants theme are analyzed in this study.

Moore and Shellman (2004) in their article analyze the exposure of the bilateral migration to the politics. They claim that the forced migration occurs when the human rights and liberties in the country of origin are violated. The results of their study show that the democracy, rule of law and the absence of the terrorism in a government decrease the amount of forced migration by 3%. Almost the same idea is stated in the article of Poprawe (2005), who claims that the volume of migration rate depends on the level of corruption regulations. The results are obtained through the use of a bilateral dataset consisting of cross-sectional data for 230 countries. According to the results of the regressions for a 2000 year, the increase of the Corruption index in the home country by 1 unit leads to the rise of the number of emigrants from this country by 30%.

According to some articles, the significance of the political reasons that serves as a push factors for emigration can depend on the level of migrant's education. Cooray and Schneid (2015) study the OECD countries emigration of low-educated, middle-educated and high-educated people from 1980 to 2010. It turns out that the rise of the corruption by 1 unit results in the 0,25% increase of emigration rate of highly-educated people, 0,23% of medium-educated people and 0,15% increase of low-educated migrants. Also, the authors conclude that the sufficiently high inequality gap between people leads to the decrease of the low- and medium- educated migrants that leave the country of origin. Foadi (2006) and Dimant, Kriege and Meierrieks (2013) come up with the idea that corruption plays a role of the push factor for the high-educated migrants that are looking for the job opportunities or better standards of living. The increase of corruption in a country results in the decrease of the education returns which forces educated migrants to leave the country of origin. Ariu and Squicciarini (2013) suggest that the lack of corruption takes on the role of the pull factor. They claim that the migrants with a high level of education would try to enter the foreign governments with a stable and favorable political situation in order to find a job in each sphere without difficulties connected with the under-the-counter money and lack of fair competition.

Some authors consider that the absence of political stability can affect the immigration and emigration rates. Veronico (2018) and Karemera, Oguledo and Davis (2000) state that the number of migrants that enter the foreign country decreases when the political instability in the country of origin picks up momentum. Lapshyna (2004) showes that the probability of emigration flows for those countries where the corruption activities are badly regulated is higher by 8%. In addition, the lack of political regulations can result in the spread of slave-trade which is considered as a forced emigration (Zhang, Pineda, 2008).

Although, a big amount of literature is dedicated to the opinion that corruption either has a negative effect on the number of migrants that enter the government or positive effect on the number of migrants that leave the country, there exist some articles that demolish that views.

According to Radnitz (2007), there is almost no effect of political factors on the migration flows. This statement was made on the basis of the interviewing 200 adults. There were 23 questions connected with the reasons of migration, occasions of human rights mistreatment, absence of democracy, life satisfaction and other problems. Radnitz (2007) concludes that the main reasons for migration are the better living conditions, job opportunities, unemployment and other factors that are indirectly influenced by politics.

In addition to Radnitz (2007) opinion that there is no effect of corruption on migration, Mckinzie (2005) suggests that the Control of corruption indirectly increases the number of people leaving the country. According to Mckinzie (2005), high passport costs can signalize an existence of a corruption activity, bad state participation in citizens problems and abuse of human rights and freedoms. Mckinzie states that the bigger are the passport costs for documents processing the smaller would be a migration flow from the country of origin to the country of destination. To prove this hypothesis Mckinzie (2005) make a regression with a dependent variable-log of passport costs and independent variable-control of corruption. The results of this regressions show that there is a negative effect of the Control of corruption on passport costs. So, the bigger is the Control of corruption, the smaller are the passport costs. In order to check the effect of passport costs on migration Mckinzie (2005) uses a bilateral migration dataset for 226 countries which included both legal and illegal migration. It turned out that when the passport costs decrease by 1% the amount of people leaving a country of origin increases by 0.54%. So, Mckinzie (2005) concludes that the better the corruption regulations are the bigger would be an outflow of migrants.

On the basis of the foregoing, we can infer that the number of articles concentrated on the corruption effect on the bilateral migration is quite limited. Only Veronico (2018), Poprawe (2005) and Mckinzie (2005) make a research using the bilateral data. It should be mentioned that there are the clashing ideas about the impact of corruption on migration flows. Also, none of the analyzed authors in their articles studies a regional migration by income. So, in order to get a full picture of the migration process and its being influenced by the political indexes both gender migration and migration from countries of different wealth are analyzed in this work.

As for the purpose of this study, we make a hypothesis that there exists an effect of corruption on migration either for the country of origin or the country of destination. We suppose that the existence of corruption and violation of human political rights would dissuade people from migration to such government. At the same time, if there exist an effect of corruption in the country of origin, then it would negatively affect people due to the fact that citizens would try to avoid such living conditions and migrate to the freer in terms of political rights government. It should be mentioned that both effects are possible. We also suppose that the effect of corruption would vary with the type of wealth of the country of origin and the country of destination.

This study consists of four parts. Firstly, we describe the data that we use, the expected results and effects. Secondly, we make the econometric models in order to evaluate the effect of corruption in the country of origin and destination for the migration from all to all countries. Then we analyze the corruption effect on bilateral migration from different income regions. Finally, we made the conclusion about possible benefits for the government authorities.

CHAPTER 2. EMPERICAL RESULTS

2.1 Data

The main variable- the number of migrants was taken from the United Nations Organization which collects the worldwide bilateral migration data from 1990 to 2019 years with 5-year intervals. The data for the Corruption Perceptions index (CPI) starts from 1995, so in our research we used the data from 1995 to 2015 years with 5-year intervals. Finally, the dataset involved the data for 165 countries for 1995, 2000, 2005, 2010 and 2015 years. The classification of the countries by their type of income (low-income, medium-income and high-income countries) was taken from the World Bank. According to the 2018 World Bank classification, the low-income countries are those ones with the GDP per capita less than $996 per year. The GDP per capita of people living in the lower-middle income countries varies from $997 to $3995. For the upper-middle income countries the GDP per capita lays in the interval from $3996 to $12055. The high-income countries have the GDP per capita more than $12055 per year. For our study we combined the lower-middle-income and upper-middle-income categories into middle-income.

As it was told before, one of the tree main indexes which are used in this work in order to check the effect of the corruption on migration is the Corruption Perceptions index (CPI). This index reflects the corruptions perceptions both of the experts and the ordinary citizens. The second one index that reflects the corruption is the Control of corruption index which states for the control of the level of the government resources or powers that were used for private purposes. The third one is the Rule of law index which is used in our work in order to look at the effect of government regulations, that include the absence of political rules violation and corruption, on the migration flow. These three indexes would allow us to fully analyze the possible existence between corruption activity in the countries of origin and destination and the number of migrants between these countries. In addition, the usage of three corruption indexes serves as a robustness check.

First of all, we need to understand whether there is a correlation between migration and corruption and law indexes, such as the CPI, Control of Corruption index and the Rule of law. The two-way scatter plots were used for this goal.

Figure 1. Correlation of log_migrants and CPI_d

Figure 2. Correlation of log_migrants and Control of Corr_d

Figure3. Correlation of log_migrants and Rule of law_d

From Figure 1- Figure 3 we can observe an existence of positive relationship between logarithm of the number of migrants & CPI in the country of destination, logarithm of the number of migrants & Control of Corruption in the country of destination and logarithm of the number of migrants & Rule of law in the country of destination.

Figure 4. Correlation of log_migrants and CPI_o

Figure 5. Correlation of log_migrants and Control of Corr_o

Figure 6. Correlation of log_migrants and Rule of law_o

From Figures 4-6 we see that there is almost no relationship between the logarithm of the number of migrants & CPI in the country of origin, logarithm of the number of migrants & Control of Corruption in the country of origin and logarithm of the number of migrants & Rule of law in the country of origin.

According to the summary statistics (A .1), we can see that the mean values of CPI, Control of corruption and Rule of law indexes are bigger for the countries of destination (“d”) than for the countries of origin. In addition, as we conclude from the graphs of correlations, there is a positive correlation between all corruption indexes in the country of destination (“d”) and the number of migrants. Also, for the country of origin (“o”) the correlation is almost invisible. So, we suppose that there would be the similar results in our bilateral regressions and the effect of corruption and law would be observed for the countries of destination.

The description and the expected signs of the dependent, independent and control variables are presented in the table below:

Table 1. Description of the variables

As it is a gravity model, all of the variables ,except the common regressors, are taken both for the country of origin and for the country of destination.

2.2 Methodology

According to the economic theory, we suppose that the most suitable method for our bilateral data is PPML which is used for gravity models (Docquier, Lodigiani, Rapoport & Schiff, 2016). The problem of bilateral data is in the possibility of the existence of zeroes in the data for migrants from one country to another. In order to avoid this problem and still interpret the changes of the dependent variable in % we replaced 0 migrants by 1 migrant. So, this step would allow us to take into account all countries migration without removing 0 migration because of taking the logarithm. In addition, the PPML method would allow us to avoid the possible heteroskedasticity.

As a result, our final model specification looks like:

where “d” stands for the country of destination and “o” stands for the country of origin; stands for the number of migrants from country “o” to country “d”; stands for the common for the country of origin and destination parameters, such as the destination between these two countries, the existence of a common culture, religion or currency; stands for the features of the country of origin, such as the GDP per capita in the country of origin or the amount of Population; stands for the features of the country of destination; is an error.

2.3 Results and Discussions

As we established before, the PPML method would be used in the final regression modelling after the choice of the best model for each of 3 indexes with the help of Random Effects regressions. The analysis of the effect of corruption on migration began from the overall analysis of migration from all to all countries around the world. We also analyzed the gender migration for these countries.

For each index four models with the addition of different independent variables were built (Table 2, A. 2, A. 3). In all of these models we can see that the higher levels of CPI, Control of corruption and Rule of law in the country of destination increase the number of migrants.

We started with the general model consisting of one of the political regressors (CPI, Control of corruption and Rule of law) and the main control variables, which are supposed to be very important factors for migrants (Poprawe, 2005). Firstly, we used the ordinary Random Effects method for regression modelling in order to have a possibility to include the variables that do not change over time. The control variables for the (I) model included: the GDP per capita, the level of Unemployment, the period (1995, 2000, 2005, 2010 and 2015 year dummies) and the main gravity model regressors which should be included in the model, such as the amount of Population in the country of origin and destination, the distance between countries, the existence of common border, common language, common religion and common currency. The period of migration was also included. Then we added the Death rate, Military and Government expenditures (II). As we can see, the results were almost the same as in the (I) model. The next step was to add the School enrollment and the index of the Shadow economy (III). Finally, we included the amount of Urban population (IV). It can be seen that almost all additional control variables turn out to be significant.

Table 2. The effect of CPI on migration from all to all countries for all migrants; RE regression

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

The final model that was chosen for our further analysis is the model IV.

The same manipulations were done with the regressions for the indexes of the Control of corruption and the Rule of law (A. 2, A. 3). Again, the IV model, which includes the most important for a migrants' decision variables, was chosen for each index. The Table 3 consists of three final models for CPI, Control of corruption and Rule of law indexes for the migrants from all to all countries. This table presents the results which were obtained using the Random Effects models. All of the models were checked for the existence of multicollinearity which was rejected (A. 6., A. 7., A. 8).

Table 3. The effect of all political indexes on migration from all to all countries for all migrants; RE regression

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

As we established before, the previous results were observed for the Random Effects regressions. Now we can use the PPML method for heteroskedasticity avoidance. The results of the PPML regressions for all 3 indexes for migration from all to all countries are presented in the Table 4.

Table 4. The effect of all indexes on migration from all to all countries for all migrants; PPML regression

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

We can see that when we do not take into account the region of immigration (entering the foreign government of destination) and emigration (leaving the home country of origin) there is a positive effect of CPI in the country of destination (CPI_d) on the number of migrants and no effect of CPI in the country of origin (CPI_o) (Table 4). It means that people in general do not care about the level of corruption in a country of origin and they leave their home country not because of political reasons. However, when they move to the foreign government, they try to choose less-corrupted and more law-regulated country. Also, it should be mentioned that people take into account common factors for the country of origin and the country of destination. Namely, the migrants would rather choose the country of destination which is closer to their home country. In addition, the existence of common border, common language and culture also increase the number of migrants between two countries. Almost the same results are observed for the Rule of law index (Table 4, regression III). As a result, when the level of CPI increases by 1 unit (corruption decreases), the number of immigrants into this country rises by 5,37% for the whole sample and by 11% after the Rule of law increases by 1 unit. As for the Control of corruption (Table 4, regression II), this factor is significant both for the emigrants and immigrants. However, for the immigrants this factor is more significant and 1 unit-increase in the country of destination accounts for 10,6% increase of immigrants into this foreign government while the 1 unit-increase in the home country results in the 1,57% decrease of people leaving this government.

It needs to be said that besides the level of corruption and law regulation in a country of destination for migrants it is also important to enter the government with a high level of GDP per capita, the existence of the Shadow economy and sufficient Government expenditures. These factors would allow them to get better standards of living and bigger amount of money. For the regression including migration from all to all countries the main parameter effecting the migrant decision about leaving the country of origin is the rate of Unemployment and Shadow economy, so people predominantly migrate in order to find a job or to be move involved into illegal business in order to derive enrichment when we consider the whole sample.

We can see that the Shadow economy is an important factor for immigrants. For instance, when the index of Shadow economy in the country of destination (Shadow econ_d) increases by 1 unit the number of migrants in the regression for CPI increases by 0,497% (Table 4, regression I), in the regression for Control of corruption increases by 0,466% (Table 4, regression II) and in the regression for Rule of law by 0,476% (Table 4, regression III). Shadow economy can be attractive for migrants due to the opportunity to take part in the illegal business where they can get a lot of money.

The next step was to estimate the gender migration from all to all countries and define the gender for which the CPI, Control of corruption and Rule of law in the country of destination are more important. The results of gender migration are presented in the Table 5.

Table 5. The effect of all indexes on migration from all to all countries by gender; PPML regression

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

We can see that for all 3 political indexes their effect in the country of destination is higher for females. However, the significance of the coefficients of CPI_d, Control of Corr_d and Rule of law_d (indexes in the country of destination) has the same strength for both genders. When the index of CPI_d increases by 1 unit the number of female migrants increase by 5,99% (Table 5, regression I) while the number of male migrants increases by 5,66% (Table 5, regression II); when the index of Control of Corr_d increases by 1 unit the number of female migrants increase by 11,9% (Table 5, regression III) while the number of male migrants increases by 11,1% (Table 5, regression IV); when the index of Rule of law_d increases by 1 unit the number of female migrants increase by 12,7% (Table 5, regression V) while the number of male migrants increases by 7,25% (Table 5, regression VI).

It should be mentioned that for male migrants the Unemployment rate both in the country of origin and destination is more important than for females. This can be explained by the fact that males go abroad in order to get some money for their families more often than females.

As for the CPI, Control of corruption and Rule of law indexes in the country of origin, they are still out of primary importance, as in the regression for migrants from all to all countries.

2.4 Country-specific effects

As we revealed, there is an effect of CPI, Control of Corruption and Rule of law in the country of destination on worldwide migration. However, from the economic theory and previous regressions it can be seen that the GDP per capita has a high impact on migration too. For example, in the regression with the CPI for the whole sample (Table 4, regression I) the increase of GDP per capita in the country of destination (GDP_d) by 1% leads to 0,11% increase of people entering this government. So, this factor is especially important for immigrants who move to the foreign country in order to find better living conditions and higher standards of living.

According to the above stated fact and to the idea that migrants tend to move predominantly to the richer countries with bigger amount of resources and working places we decided to analyze the migration from the whole world to high-income countries (Ritchey, 1976). Our goal is to check whether the effect of CPI_d, Control of Corr_d and the Rule of law_d will stay if we take into account that migrants move to the countries with the high level of GDP which can take upon itself the effect from CPI, Control of Corruption and Rule of law.

From all countries to high-income countries

Table 6. The effect of corruption on migration from all to high-income countries; PPML regression

Robust standard errors in parentheses

*** p<0.01, ** p<0.05, * p<0.1

Now we can see the results of migration from all to high-income countries: 2,41% increase of migrants into the country of destination for 1 unit CPI_d increase (Table 6, regression I), 2,79% increase of migrants into the country of destination for 1 unit Control of Corr_d increase (Table 6, regression II) and 0,677% increase of migrants into the country of destination for 1 unit Rule of law_d increase (Table 6, regression III). Also, when people migrate to high-income countries the Unemployment in the country of origin becomes much more significant for them than it was in the regression for world-wide migration. The value of GDP_d per capita for the migration to high-income countries became bigger in comparison with the coefficient of the GDP_d in the regression for migrants from all to all countries (Table 4), which proves the hypothesis about the willingness of people to migrate into more rich countries because of money (there was a 0,111% increase of migrants after the increase of GDP_d per capita on 1% for the regression with CPI, 0,0813% for the regression with the Control of Corruption and 0,0722% increase for the Rule of law regression; now it is 0,244% increase, 0,178% increase and 0,181% increase). Despite the fact that the significance of GDP increased, the significant of CPI, Control of corruption and Rule of law in the country of destination still exists and is not displaced by GDP per capita.