Tyrants and Migrants

Abstract

This article examines the determinants of immigration policy toward low-skilled workers across 13 relatively wealthy autocracies after World War II. I argue that authoritarian immigration policy is a consequence of an autocrat’s redistributive policy. As the distribution of resource rents in rentier autocracies reduces the incentive of domestic labor to enter the labor force, rentier states rely on migrant workers to meet the demand for low-skilled labor. Autocrats without resource rents, however, lack capacity for redistribution, so they use policies that provide people with wages in exchange for their labor while restricting immigration. Using a policy index that measures the extent to which low-skilled migrant workers can get into a country in a given year, I find strong evidence for this argument across 13 autocracies in the post-World War II era.

Keywords

migration political economy non-democratic regimes

Introduction

Much of the political science literature on the politics of immigration focuses on the West. Foreign nationals seeking liberal political institutions and better economic opportunities are drawn to wealthy democracies around the world. Given the political, economic, and demographic significance of immigration in developed democracies as well as the emergence of right-wing populism and fiscal crises in Europe, political scientists have explored the roles of party systems, welfare policy, xenophobia, and special interests in shaping immigration policy developments under democratic institutions.¹ Only a handful of studies ask questions about the causes and consequences of immigration in autocracies, while few studies attempt to explain immigration policy variation in wealthy autocracies, such as Saudi Arabia and Singapore.²

A number of factors explain why the literature has done little to fill this gap. First, the literature on the politics of immigration evolved in isolation from the literature on the politics of authoritarianism. The political science literature on immigration is based on decades of research spearheaded by economists and demographers focusing on advanced democracies. Many of the assumptions and theories are not applicable to institutional settings of authoritarian regimes. Second, the lack of data on authoritarian governments’ immigration policies has been a significant setback to any quantitative attempt to identify determinants of immigration policy in autocracies.³ Finally, few studies have considered immigration policies of autocracies in comparative perspective. While scholars of the Middle East and the students of the rentier state note the importance of immigration in the Persian Gulf, their analysis of immigration is often descriptive or limited to the region.

The Low-Skill Immigration Policy Dataset of wealthy autocracies allows this article to explain why some authoritarian governments open immigration while others restrict immigration inflows and why an autocracy’s policy stance on immigration changes over time. In addition, the dataset of this article builds on an existing dataset to perform a more comprehensive factor analysis and adds more observations to the sample used in the empirical section. Using the actual laws, enforcement mechanisms, and other de facto policies concerning immigration, the policy index seeks to capture 13 autocracies’ policy shifts over time after World War II (WWII).

Autocracies’ divergent immigration policies have resulted in different population shares of immigrants. The presence of mass immigration or lack thereof in wealthy autocracies suggests a new venue of research in our understanding of the politics of authoritarian regime and autocratic survival. More precisely, does immigration help authoritarian governments tighten their grip on power? And given the possible political and economic effects of immigration on authoritarian persistence, why do some autocracies have more restrictive immigration policy than others? This article primarily addresses the latter question by looking at a number of high-growth, relatively wealthy authoritarian countries.

Economic growth increases the demand for labor, especially low-skilled workers. Given the high labor demand, why do some authoritarian governments seek foreign labor while others meet the labor demand by encouraging more native citizens to participate in the labor force? I primarily focus on relatively wealthy autocracies because one of the most important causes of cross-border migration is the wage differential between migrant-receiving and migrant-sending states (Abella, 1995; Breunig, Cao, & Luedtke, 2012; Hanson & Spilimbergo, 1999; Massey et al., 1993; Ortega & Peri, 2013).⁴ As poor autocracies do not face significant immigration pressure, they can set up immigration policy that lacks enforcement and effectiveness. Moreover, it would be difficult to code and make inferences about poor economies’ immigration policy. High economic growth and economic prosperity of some autocracies suggest that immigration policies of these authoritarian governments are relevant and deserve scholarly attention.

I argue that the immigration policy of an authoritarian regime is a consequence of elites’ redistributive policy and their concern about the labor market. Even though elites generally prefer immigrant labor, immigration policy depends on the extent to which autocrats are able to redistribute to native citizens who would be underemployed in the presence of substantial low-skill immigration. When governments rely solely on elites’ tax revenues, they lack capacity to redistribute. Without redistribution, autocrats provide wages in exchange for native workers’ labor. Revenue-seeking governments encourage labor-market participation of domestic workers while restricting immigration. However, governments with independent sources of income distribute rents to their citizens while supplying migrant workers to elites who support the regime.

This article illustrates how autocratic elites use immigration policy to meet their specific political goals, given their redistributive policies and political constraints. Hence, it is helpful to understand why immigration policies differ across autocracies before one can assess the consequences of immigration under authoritarian regimes. Using the existing literatures on the political economy of authoritarianism, rentier state, and redistribution, the article focuses on autocrats’ discretion over resource income as one of the key causes of immigration policy in wealthy autocracies.⁵

First, I begin by theorizing how an autocracy’s distribution of resource rents leads to open immigration policy through its effect on redistribution and natives’ incentive to enter the labor force. Then, I provide an overview of authoritarian immigration policy by comparing autocracies’ immigration policy dimensions to those of democracies. Third, I provide evidence that differences in natural resource income help explain why some autocrats open immigration while others do not make policy changes to attract immigrants or even discourage immigration. Fourth, I use an alternative measure of immigration policy openness and conduct additional robustness checks. Finally, I conclude and discuss the implications of immigration within the broader political science literature of authoritarianism and introduce new areas of research in the field.

Politics of Authoritarian Immigration Policy

Common explanations for immigration policy formation invoke a number of cultural, economic, and sociotropic origins of native opposition to low-skill immigration. Xenophobia, labor unions, and citizens’ concerns for taxation, inequality, and welfare policy are assumed to influence immigration policy because elections punish incumbents whose policies do not reflect popular preferences over immigration. Scholars studying immigration policies of representative democracies have examined whether firms influence immigration policy by supplying financial resources to policymakers.⁶ While scholars disagree on the extent to which citizens in democracies influence immigration policy, immigration tends to be a highly salient issue about which voters have strong opinions. In representative democracies, both average citizens and special interests play roles in shaping government policy (Grossman & Helpman, 1994, 2001). Reelection-minded policymakers in democracies must strike a balance between pro-immigration businesses and voters. As authoritarian rulers are largely insulated from mass preferences over government policy, their policy choices primarily reflect their concern for regime stability (Bueno de Mesquita, Smith, Siverson, & Morrow, 2004).

A central puzzle of this article concerns what influences elites’ immigration policy choices.⁷ Why do some authoritarian governments of fast-growing economies become increasingly dependent on foreign labor, while others limit immigration? Moreover, what explains an authoritarian government’s policy shifts over time? Economic needs for low-skilled labor do not explain why oil-rich autocracies in the Persian Gulf have relied on immigration more than the export-oriented economies of East Asia. The so-called East Asian Tigers experienced exponential growth under dictatorships, driven by booming labor-intensive manufacturing sectors. The growth rates of these export-oriented economies often exceeded those of the Persian Gulf fueled by petroleum exports, a capital-intensive industry that requires relatively little labor. Given the sectoral differences between the oil industry and manufacturing sectors, economic theories predict that the East Asian Tigers should have relied on more foreign labor than the oil-rich autocracies in the Middle East.

Domestic Interests Over Redistribution and Labor Demand

This section illustrates the policy preferences of autocratic elites and native citizens over redistribution and immigration policy. For simplicity, I do not make a distinction between elites and autocratic rulers. I instead assume that both elites and autocratic rulers have vested interests in regime stability through an intimate government–business alliance. I relax this assumption later to see how economic elites’ distinct preferences for redistribution and immigration shape autocrats’ immigration policy. I discuss two interrelated channels through which an increase in autocrats’ resource rents leads to more open immigration policy: (a) the effect of non-tax revenue on the labor market and (b) the effect of non-tax revenue on autocrats’ vulnerability to native citizens’ anti-immigrant attitudes.

Autocratic elites and native citizens have opposing preferences for redistribution and immigrant labor. While elites seek to limit redistribution and have a strong preference for foreign labor, native citizens favor redistribution and oppose foreign labor.⁸ The core motivation behind elites’ opposition to redistribution is based on the assumption that elites have to pay for native citizens’ welfare. However, native workers oppose foreign labor because competition with immigrant workers decreases their income and on other non-material grounds (Freeman, 1995; Hatton & Williamson, 2005, 2007; Zolberg, 1989).⁹ However, substantial redistribution can keep domestic labor quiescent about open immigration policy because income transfers decrease native citizens’ incentive to enter the labor force and the extent to which governments are susceptible to their citizens’ anti-immigrant attitudes. Even though elites in autocracies are relatively insulated from popular opposition to foreign labor, immigration policy outcomes depend on the extent to which they can use redistribution and the effect of redistribution on native citizens’ incentive to work and to oppose immigration.

Autocrats can rely on various tools to consolidate their power, both coercive (i.e., repression) and persuasive (i.e., cooptation; Wintrobe, 1998). Elites who depend on native labor must master difficult strategies of wage repression and seek ways to prevent and counteract collective actions by domestic labor for economic and political reforms.¹⁰ Moreover, elites’ economic dependence on domestic labor increases labor’s influence on policymaking. Furthermore, labor may demand more policy concessions when autocrats tax labor. However, if autocrats can rely on temporary migrant workers from poorer countries, they can prevent political complication arising from native workers’ employment opportunities. As temporary migrant workers have a short time horizon in host autocracies and are primarily attracted by the wage differential between sending and receiving economies, they are less likely to protest and risk repression for labor and political reforms.¹¹

While elites have a strong preference for access to foreign labor, they face different domestic constraints when making labor-market and immigration policies. The extent to which elites open immigration depends on two factors. First, the greater the extent to which elites can send direct income transfers to native citizens, the more foreign labor they can bring. Second, the source of income transfer matters. When elites have to carve out their earned income for redistribution, they are less likely to redistribute. In contrast, elites with access to natural resource rents have the means to send government transfers in exchange for loyalty.¹² Morrison (2009) finds that an increase in non-tax revenue is associated with more social spending in autocracies. Without redistribution to underemployed domestic labor, autocrats who open immigration, however, face increasing popular pressure for regime change.¹³

To summarize the argument, elites’ distribution of natural resource windfalls to their citizens decreases citizens’ incentives to take low-skilled jobs and to stage opposition toward temporary migrant workers. The transfer of rents to citizens, therefore, increases both elites’ incentive and ability to open immigration. As native citizens become more expensive to hire as a consequence of redistribution, autocrats bring foreign labor to meet the labor demand. Elites who lack capacity to redistribute, however, use native labor to meet the labor demand while engaging in other labor-market policies to increase their profit and to stay in power. The combination of immigration and redistribution exempts elites from engaging in repressive labor-market policies toward domestic labor and becoming accountable to citizens’ immigration policy preferences. Therefore, I argue that natural resource income is positively correlated with immigration policy openness in autocracies. By contrast, the lack of resource income characterized by the East Asian Tigers under authoritarianism should lead to restrictions on immigration. Thus, I hypothesize the following:

Hypothesis 1: As the size of an autocrat’s resource income increases, the openness of immigration policy increases.

Autocrats, Elites, and Economic Policies

For certain autocracies with few state-owned enterprises, it may be more useful to think of economic elites and autocratic rulers as separate entities with different goals. According to this framework, autocratic rulers only care about staying in power while economic elites are primarily concerned about maximizing profit. Revenue-seeking governments in non-rentier states depend on taxes generated from economic activities of both capital and labor (Bates & Lien, 1985). Without redistribution, autocrats of growing economies must provide work opportunities for native citizens to prevent popular dissent.

Knowing the policy constraints of autocrats, economic elites seek two policy concessions from authoritarian regimes. First, elites oppose redistribution of their profit to labor. Second, instead of foreign labor, elites seek government-sponsored wage repression and other policies that discourage strikes and the mobilization of workers. Although the cost of repression is high when citizens have workplaces to overcome collective action problems (Boix, 2003), capitalists pass on the cost of repression to autocrats. Revenue-seeking governments relying on capital’s tax contribution accommodate capital’s preferences in their social spending and labor-market policies. The East Asian welfare model in 1980 exemplifies this framework in a non-rentier economy. The flexible labor market and limited public provision of social insurance outside of state sector workers in East Asia are examples of wage repression and limited redistribution (Haggard & Kaufman, 2008). However, governments in East Asia emphasized the importance of basic health services and education as public health and education increase the productivity of labor.

The presence of non-tax revenue, however, blurs the distinction between autocrats and economic elites because authoritarian governments in rentier states collect revenues from state-owned enterprises (Beblawi, 1987; Mahdavy, 1970). According to this view, autocrats are economic elites who are primarily interested in maintaining regime stability. Autocrats’ ability to generate profit (e.g., resource rents) also translates into their ability to buy off opposition and disperse popular dissent through direct income transfers. As the distribution of resource rents makes native citizens wealthier and less opposed to immigration, autocrats rely on foreign labor to meet the labor demand.

Overview of Authoritarian Immigration Policy

While many theories of immigration policy formation have emerged across disciplines, the lack of cross-sectional time-series data on immigration policy has kept the discipline from performing rigorous statistical evaluation of the theories. Moreover, we do not know whether and how immigration policies differ across various political institutions. In this section, I describe how I augmented the recently available dataset constructed by Peters (2015). I also present descriptive statistics on the 15 immigration policy dimensions by regime type.

Factor Analysis

The immigration policy variable is a factor score based on 12 dimensions of immigration openness, listed in Table 1. Each dimension takes a score ranging from 1 to 5, with the latter indicating a more liberal policy stance toward immigrants. The final factor score covers a variety of immigration regulations and laws that seek to control immigration flows by screening potential immigrants. The final score generally excludes the dimensions that do not concern the labor-market and business interests, such as policies concerning refugees and family reunion. While most scores come from actual immigration laws in effect, executive policy discretion over deportation and enforcement contributes to the final makeup of the factor score. In sum, the factor score measures how many low-skilled workers can enter a country in a given year.

Table 1.

Dimensions of Immigration Policy.

Dimension	Description
Universality by Nationality	Entry discrimination based on nationality
Universality by Skill or Income	Entry discrimination based on skills or income
Citizenship	Ease of naturalization or citizenship acquisition
Immigrant Rights	Political, legal, or welfare rights
Refugee	Number of refugees allowed to enter
Asylum	Ease of getting an asylum
Recruitment	Visas or government programs
Labor Prohibitions	Labor-market restrictions for immigrants
Deportation	Deportable offenses and administrative processes
Enforcement	Border enforcement or employment screening
Family	Sponsorship by citizenship and restrictions
Quota	Percentage of population allowed to enter annually

See Online Appendix B for details and examples of the coding scheme and each dimension of immigration policy.

The original dataset in Peters (2015) includes 19 democracies and autocracies from the late 18th century to 2010, covering up to 225 years. I add 10 more countries to the dataset for the purpose of a more comprehensive factor analysis by using the codebook in Peters (2015). Then, I use a factor analysis based on principal components to construct an immigration policy index by pooling all 29 countries, including both democracies and autocracies and all years from 1783 to 2013. Although the theory of this article is only applicable to a sample of autocracies, the theoretical justifications for sample selection should not decide which observations be included to generate the immigration policy index in the factor analysis. Moreover, it is my goal to take advantage of as many observations of immigration policy dimensions as possible to recover a representative immigration policy score. Furthermore, I do not see any fundamental coding discrepancies between immigration policies of democracies and autocracies based on the codebook.

The factor analysis measures the restrictiveness of immigration policy by retaining two factors. The first factor is primarily derived from policy dimensions that control immigration flows by erecting legal or administrative barriers to potential new immigrants. However, the second factor puts more weight on various social provisions and restrictions to which immigrants are subjected as residents of receiving states. The factor loadings and scoring coefficients in Table 2 indicate that the conceptualization of Factor 1 as Immigration Policy and Factor 2 as Immigrant Rights is justifiable.

Table 2.

Factor Loadings and Scoring Coefficients.

Variable	Factor 1		Factor 2		Uniqueness
	Immigration policy		Immigrant rights
	Factor loadings	Scoring coefficients	Factor loadings	Scoring coefficients
Nationality	0.486	0.124	−0.089	−0.043	0.411
Skill	0.737	0.187	0.047	0.023	0.420
Citizenship	0.197	0.050	0.598	0.291	0.531
Other Rights	0.378	0.096	0.724	0.352	0.313
Refugees	−0.670	−0.170	0.429	0.209	0.340
Asylum	−0.556	−0.141	0.435	0.212	0.431
Recruitment	0.548	0.139	0.091	0.044	0.503
Work Prohibitions	0.450	0.115	0.566	0.275	0.439
Deportation	0.600	0.153	0.472	0.229	0.358
Enforcement	0.763	0.194	−0.018	−0.009	0.357
Family Reunification	−0.672	−0.171	0.355	0.173	0.371
Quota	0.561	0.143	−0.337	−0.164	0.481

All values are rounded to three decimal places.

The immigration policy score now ranges from −2.95 (most restrictive) to 1.52 (most liberal) with 0 mean and unit variance in the whole sample. The score correlates highly (at 0.942) with a standardized average of nationality, skill, quota, recruitment, labor prohibitions, deportation, and enforcement.¹⁴ Peters (2015) shows strong evidence that the immigration policy variable is positively associated with immigration inflows with a 1 standard deviation increase leading to 1.43 more immigrants per million U.S. dollars of gross domestic product (GDP).

Immigration Policies by Regime Type

I present descriptive statistics on immigration policies across different political institutions after WWII. I make a distinction between democracies and autocracies by using the regime dataset provided by Przeworski, Alvarez, Cheibub, and Limongi (2000) and Cheibub, Gandhi, and Vreeland (2010).¹⁵ Although the authors do not identify Hong Kong’s regime type, it can be considered a liberal autocracy during its time as a British colony with civil freedom but without representation (Ma, 2007). The same can be said about Hong Kong as a Chinese territory. With Hong Kong, Table 3 lists 13 autocracies.

Table 3.

Autocracies Included in the Sample.

Group	Country	Years of autocracy
Settler States	Argentina	1955-1982
	Brazil	1964-1978
	Chile	1973-1989
	South Africa	1950-1993
	Venezuela	1950-1958, 2009-2012
Asian Tigers	Singapore	1960-2010
	South Korea	1948-1987
	Taiwan	1951-1995
	Hong Kong	1966-2010
Oil-rich Monarchies	Saudi Arabia	1950-2010
Oil-rich Monarchies	Kuwait	1963-2010
Other Autocracies	Spain	1950-1976
Other Autocracies	Botswana	1966-2013

Argentina, Brazil, Chile, South Africa, South Korea, Taiwan, and Venezuela have democratized in various years. The sample only includes country-year observations under authoritarian regimes. Some countries are included after 1950 due to missing data on immigration policy, explanatory variables, or controls.

Although one can expand the dataset by coding the immigration policies of other autocracies that are popular destinations of international migrants, such as Qatar, the United Arab Emirates, and Bahrain, it is extremely difficult to obtain primary and secondary sources on immigration policies of these countries. In addition, adding more rentier states to the dataset will make the sample unbalanced. In such a sample, any empirical support from the dataset may arise from the dataset’s heavy reliance on oil-rich monarchies. As the sample includes at least two countries from South America, Africa, Asia, and the Middle East, the dataset represents a variety of autocracies around the world.

In general, autocracies place fewer restrictions on immigration inflows than democracies.¹⁶ This is evident in Universality by Nationality, Quota, and Enforcement shown in Table 4. Autocracies’ high scores in these dimensions imply that immigrants are better able to find entry into autocracies. Autocracies, however, limit foreigners’ access to citizenship and rights as well as family reunion, specifically the extent to which migrant workers can bring family members. This is consistent with a recent finding that autocracies can absorb more immigration than democracies (Breunig et al., 2012). The trade-off between entry policies and rights provisions also resonates with a notion that high-income countries cannot increase both immigration openness regarding entry and the extent to which migrant workers are granted rights after admission (Ruhs, 2013).

Table 4.

Immigration Policy Dimensions by Regime Type, 1946-2013.

Dimension	Democracies			Autocracies
Dimension	M	SD	Observations	M	SD	Observations
Universality by Nationality	4.014	1.275	1,318	4.318	0.911	504
Universality by Skill or Income	3.043	1.05	1,318	3.106	1.15	504
Citizenship	3.815	0.866	1,318	3.197	1.027	504
Rights	3.887	0.883	1,318	3.176	1.01	504
Refugee	2.507	1.24	1,318	1.236	0.626	504
Asylum	2.715	0.989	1,318	1.333	0.842	504
Recruitment	3.192	1.035	1,318	2.873	1.103	504
Work Prohibitions	4.134	0.959	1,318	3.849	1.098	504
Deportation	2.889	0.855	1,318	2.389	0.942	504
Enforcement	3.135	0.882	1,318	3.712	0.847	504
Family	2.643	1.15	1,315	2.045	0.828	504
Quota	4.225	1.435	1,318	4.776	0.767	504
Refugee Provision	0.780	0.414	1,318	0.367	0.482	504
Asylum Provision	0.869	0.337	1,318	0.353	0.478	504
Family Provision	0.937	0.243	1,318	0.940	0.237	504
Immigration Policy	−0.856	0.75	1,315	−0.387	0.44	504

The most striking policy divergence between democracies and autocracies, however, is shown in refugee and asylum policies. If provisions regarding refugees and asylees exist, Refugee Provision and Asylum Provision are coded “1” (“0” otherwise). While most democracies and autocracies have family provisions, autocracies are much more restrictive toward refugees and asylees. Autocracies’ hostility toward refugees and asylees is also evident in Refugee and Asylum. As the score of 1 means almost no refugees or asylees are allowed in host states, autocracies’ average scores of 1.236 and 1.333 for Refugee and Asylum, respectively, indicate that autocracies have extremely restrictive policies toward refugees and asylees. However, democracies’ policy scores toward refugees and asylees are close to 3, meaning that democracies allow a moderate number of refugees while following the United Nations’ definition of a refugee and procedures.

Another notable policy difference between democracies and autocracies is that autocrats’ deportation policies are designed to expedite migrant worker deportation. The score of 2 for Deportation means that there are relatively many deportable offenses for migrant workers and that authorities can use a simple administrative process to deport a migrant worker. Autocrats relying on foreign workers tend to use deportation as a convenient tool to manage the labor market. This sharply contrasts with autocracies’ more liberal entry criteria. Overall, the Immigration Policy factor score—measuring the extent to which migrant workers can gain entry into host states—is much higher for autocracies. Using the Polity IV data series with 6 as the lowest score for democracies and scores lower than 6 as autocracies shows very similar differences in immigration policies between democracies and autocracies.¹⁷

International Migration and Regime Type

Research on international migration and regime type shows that there are four patterns in the post-WWII age of migration (Breunig et al., 2012). First, migrants tend to seek better economic opportunities, not civil and political freedom. Second, migrants tend to come from relatively poor democracies. As democracies allow the freedom of cross-border emigration, citizens from democracies have a greater degree of cross-border mobility in terms of exit. Third, autocracies absorb more migrants because they are less constrained by domestic demand for anti-immigration policies. Finally, a recent work shows that autocracies’ emigration policies also vary because autocracies face different trade-offs between the costs of emigration (e.g., mass exit and foreign influence) and the benefits of emigration (e.g., expelling dissidents and receiving remittances; Miller & Peters, 2015).

The existing research on international migration and regime type makes inferences about autocracies’ immigration policies based on actual immigration inflows. The descriptive statistics of the 13 autocracies’ immigration policies in this article confirm some of the findings of the existing research, particularly the finding that autocracies implement more open immigration policy while granting few rights to migrant workers. Table 5 portrays the differences in population shares of foreign-born individuals between democracies and autocracies included in the Low-Skill Immigration Policy Dataset.

Table 5.

Migrant Shares (%) by Regime Type, 1946-2013.

Regime type	M	SD	Minimum	Maximum	Observations
Democracies	8.467	6.004	0.352	22.528	1,016
Autocracies	21.876	23.052	0.542	81.42	370
Autocracies except K&SA	15.044	16.383	0.542	52.916	276

Data on immigration are retrieved from the online World Development Indicators database by the World Bank (2014), measuring foreign-born population percentage (international migrant stock as a percentage of population). The World Bank does not provide annual observations on immigration but in increments of 5 years starting in 1960. As migration stocks tend to be stable and follow linear trends, linear interpolation is used between time points to fill missing observations. K&SA = Kuwait and Saudi Arabia.

Consistent with the existing literature on international migration and regime type, autocracies have larger population shares of migrant workers. The row, Autocracies except K&SA shows that excluding the oil-rich monarchies, Kuwait and Saudi Arabia from the pool of autocracies does not change the central finding that autocracies have more capacity to absorb foreign workers than democracies. Moreover, autocracies’ population shares of foreign labor exhibit a much higher standard deviation than those of democracies. However, democracies show a clear limitation in such variation relative to autocracies.

Empirical Analysis

Data on Natural Resource Income

To assess the relationship between immigration policy openness and natural resource income, I have considered several measures of natural resource rents. The literature on the resource curse has used revenues from natural resource exports as a share of GDP as the standard measure of resource dependence (Morrison, 2009; Ross, 2001; Smith, 2004). While this measure roughly captures the fiscal dependency of governments on natural resource rents, it ignores the actual size of resource income that governments can use for regime stability. As the main theoretical premise of the hypothesis is that autocratic elites of rentier states distribute income to silence dissent among people, a more appropriate measure of resource income is the total revenue from natural resource production divided by population. Natural resources include fuel (oil, gas, and coal) as well as valuable minerals (i.e., gold, diamond, silver, and copper). To measure the value of natural resource production, the production quantity of each resource is multiplied by its real world price, expressed in thousands of 2007 U.S. dollars. The total income from all natural resources is then divided by population in a given year. Starting with the dataset provided in Haber and Menaldo (2011), I have expanded the data on resource income for years up to 2013 by using growth rates of resource income from World Bank (2014).¹⁸

A remaining issue with the natural log of resource income per capita is that it underestimates the extent to which autocrats can provide patronage to citizens. As rentier states such as Kuwait and Saudi Arabia host large population shares of migrant workers, using resource income divided by the entire population does not precisely capture autocrats’ capacity to redistribute. Most autocrats do not generally distribute their resource income to temporary migrant workers, so I use the interpolated data of international migration stocks to compute resource income per citizen.¹⁹ For the 13 autocracies after 1945, the correlation between total resource income per capita and total resource income per citizen is .925 and the correlation between the natural logs of the two values is .997. As some autocracies, notably, Taiwan lack migration data, I replace the missing values of the natural log of resource income per citizen with the natural log of total resource income per capita.

Data on Control Variables

I select several key control variables that are critical in properly testing the relationship between resource income and immigration policy. First, I control for the 1-year-lagged value of the natural log of population. Natural resource abundance in autocracies may attract immigrants or discourage emigration, raising population which in turn has potential consequences on immigration policy. I lag population by a year to partially address the endogeneity of population to immigration policy. Second, as many resource-rich countries tend to be wealthy, I include the natural log of GDP per capita. Theoretically, differences in immigration policy between wealthy and less wealthy autocracies are plausible. Third, the amount of resource income can positively or negatively affect growth rates through its effect on the economy and on manufacturing sectors. As growth rates may change autocrats’ stance on immigration policy, I control for economic growth. Controlling for growth rates and economic wealth allows the regression to compare immigration policies of resource-rich and resource-poor economies across observations with comparable economic characteristics. Finally, when I exclude Hong Kong from the sample, I include the polity score from Marshall and Gurr (2014) to control for the level of political development.²⁰ The rentier state literature consistently finds that resource income is negatively correlated with the quality of institutions, which in turn may affect autocrats’ ability to influence immigration policy.²¹

Sample Selection

An underlying assumption behind the theoretical expectation of the relationship between natural resource rents and immigration policy openness is autocrats’ discretion over resource revenues. In the mid-1970s, many governments of rentier states nationalized their natural resource industries. As autocrats placed resource industries under their full control, their capacity to distribute rents to citizens increased as well. For instance, there is more compelling evidence for the political resource curse after 1975 (Andersen & Ross, 2013; Ross, 2012). The resource curse literature emphasizes the important role of ownership in conditioning the extent to which natural resource rents empower autocrats (Jones Luong & Weinthal, 2006, 2010).

An ideal way to select a sample is to drop all observations under which autocracies do not have ownership over natural resource industries. While Kobrin (1980) provides some information regarding when nationalizations took place and which industries were nationalized in autocracies, two problems still remain. First, the nationalization data do not cover all the years and countries in the sample of autocracies. Second, some autocrats nationalized their natural resource industries at different times during the 1970s. For instance, Argentina nationalized nine firms in the petroleum industry in 1963, followed by two more firms in 1974.²² It is difficult to operationalize the role of ownership, given the incomplete data and the complexity in the nationalization process of natural resource industries. Instead, I divide the sample into two time periods: 1946-2013 and 1975-2013.²³

Another issue is the classification of regime type. The regime dataset assembled by Przeworski et al. (2000) and Cheibub et al. (2010) classifies Botswana and Venezuela as an autocracy and a democracy, respectively, for most time periods. The quality of Venezuela’s democratic institutions started deteriorating when Hugo Chávez came into power in 1999. The polity score of Venezuela between 2006 and 2013 is below 6, placing Venezuela in a group of open anocracies according to the Polity IV Project. Botswana’s polity score, however, ranges from 6 to 8, while it is still considered an autocracy by the binary regime classification due to the lack of electoral competition. Therefore, as an alternative measure of democracy, I restrict the sample to countries with polity scores below 6.

Empirical Strategy

I regress immigration policy on the natural log of resource income per citizen, $l n$ (Resource Income per Citizen) using ordinary-least-squares (OLS) model with cluster-robust standard errors. Each model includes country fixed effects to capture country-specific, time-invariant variables that may drive immigration policy, such as culture, religion, geography, or climate. Country fixed effects also account for autocratic regime types, such as military regimes, single party, and dynastic monarchy. Year fixed effects are included to account for common shocks every year. In addition, a lagged value (1-year) of immigration policy is included to appropriately model the dynamic process as suggested by Beck and Katz (1995). As immigration policy is a set of laws and practices, the best predictor of today’s policy is yesterday’s policy. Policymakers often make amendments to existing laws and tweak existing practices to meet their goals. While some policy shifts are dramatic, past policies are the most important foundation of a country’s current immigration policy. The following model summarizes the aforementioned empirical strategy.

\begin{matrix} Immigration {Policy}_{i t} = β_{0} + β_{1} Immigration {Policy}_{i t - 1} \\ + β_{2} Natural Log of Resource Income per {Citizen}_{i t} \\ + \sum_{k = 3}^{K} (β_{k} Control {Variable}_{(k - 2), i t}) + α_{i} + μ_{t} + ε_{i t}, \end{matrix}

where $α_{i}$ and µ_t indicate country fixed effects and year fixed effects, respectively.

I run two sets of models with two different samples of autocracies. The first set of models pools autocracies based on the binary indicator of autocracy. The second set pools countries that score lower than 6 in the Polity IV Project. Each set contains four models: (a) all autocracies since 1945, (b) all autocracies without Hong Kong after 1945 and the addition of polity score as a covariate, (c) all autocracies since 1975, and (d) countries in (b) since 1975 with the addition of polity score as a covariate.

Results

Table 6 reports the regression results for the two time periods by pooling autocracies using the binary regime classification. Across the four models, the coefficient of resource income is positive. In Models 3 and 4, there is overwhelming evidence for the hypothesis. For instance, a 100% increase in resource income is correlated with a .012-unit increase in immigration policy openness in Model 3. Compared with Models 1 and 2, this correlation in Model 3 is significantly higher and highly statistically significant. The correlation between immigration policy and resource income is weaker substantively and statistically when all years since 1946 are pooled together.

Table 6.

Authoritarian Immigration Policy Regressed on Resource Income (Autocracy = 1).

	Model 1	Model 2	Model 3	Model 4
Years included	1946-2013	1946-2013	1975-2013	1975-2013
Immigration policy_t−1	0.838***	0.813***	0.700***	0.719***
Immigration policy_t−1	(0.028)	(0.033)	(0.046)	(0.056)
ln (Resource income/citizen)	0.004	0.001	0.012**	0.010***
ln (Resource income/citizen)	(0.003)	(0.002)	(0.003)	(0.001)
ln (Population)_t ₋₁	−0.119***	−0.129**	−0.340*	−0.272^†
ln (Population)_t ₋₁	(0.027)	(0.034)	(0.142)	(0.145)
ln (GDP per capita)	−0.027*	−0.031*	−0.061	−0.060
ln (GDP per capita)	(0.012)	(0.011)	(0.053)	(0.043)
GDP growth	−0.043	0.003	−0.059	−0.030
GDP growth	(0.056)	(0.031)	(0.079)	(0.085)
Polity score		0.001		−0.000
Polity score		(0.001)		(0.002)
Observations	461	413	268	231
Countries	13	12	13	12
R ²	.876	.878	.806	.817

This table portrays a pooled cross-sectional time-series ordinary-least-squares (OLS) analysis of immigration policy in year t. All independent variables are taken from year t unless otherwise noted. Cluster-robust standard errors are shown in parentheses. Country and year fixed effects are included in all models. GDP = gross domestic product.

***

Statistical significance level of 0.1%. **Statistical significance level of 1%. *Statistical significance level of 5%. ^†Statistical significance level of 10%.

It is difficult to interpret what this correlation really means in terms of policy impact due to the arbitrary nature of factor analysis. As an example, a 0.2-unit decrease in U.S. immigration policy score from 1923 to 1924 reflects the Immigration Act of 1924, also known as Johnson-Reed Act that limited the annual inflows of immigrants from any country to 2% (from 3%) of the existing U.S. residents from that country, recorded in the 1890 census (Hatton, 2010). The act aimed to cut annual immigration flows by more than 30%. Thus, it is plausible to assume that a 100% increase in resource income is associated with more than a 1.5% increase in annual immigration flows. As $l n$ (Resource Income per Citizen) ranges from 0 to 12.2 across autocracies, this is a sizable correlation.

As expected, population shows a negative correlation with immigration policy, implying that autocracies take population into account when making immigration policy. Wealthy autocracies also tend to restrict immigration, but the relationship is not statistically significant since 1975. In addition, we cannot reject the null correlation between the level of political development and immigration policy openness. Polity scores often show very little temporal variation within each country. Including fixed effects is likely to absorb any correlation between polity scores and immigration policy openness.

Table 7 reports the results for the same models as shown in Table 6 by pooling any country-year observations with polity scores lower than 6. Consistent with the results in Models 1 through 4, the correlation between resource income and immigration policy openness after 1974 (Models 7 and 8) is substantively higher with lower standard errors compared with the results in Models 5 and 6 in which all years are included. The coefficient is also all statistically significant in Models 5 and 8. The differences between Models 1 and 2 and Models 5 and 6 show that using different classifications of autocracy matters when all years since 1946 are pooled together. Using this alternative classification of autocracy lends stronger empirical support for the hypothesis.

Table 7.

Authoritarian Immigration Policy Regressed on Resource Income (Polity < 6).

	Model 5	Model 6	Model 7	Model 8
Years included	1946-2013	1946-2013	1975-2013	1975-2013
Immigration policy_t−1	0.865***	0.836***	0.738***	0.749***
Immigration policy_t−1	(0.026)	(0.033)	(0.051)	(0.063)
ln (Resource income/citizen)	0.011*	0.008*	0.016**	0.012*
ln (Resource income/citizen)	(0.004)	(0.004)	(0.004)	(0.004)
ln (Population)_t ₋₁	−0.058^†	−0.065^†	−0.214*	−0.207*
ln (Population)_t ₋₁	(0.029)	(0.033)	(0.088)	(0.074)
ln (GDP per capita)	0.012	0.009	−0.011	−0.024
ln (GDP per capita)	(0.014)	(0.016)	(0.047)	(0.044)
GDP growth	0.019	0.075	0.038	0.122
GDP growth	(0.072)	(0.048)	(0.142)	(0.144)
Polity score		0.003		−0.005
Polity score		(0.002)		(0.007)
Observations	462	418	232	196
Countries	13	12	12	11
R ²	.827	.825	.745	.754

***

Statistical significance level of 0.1%. **Statistical significance level of 1%. *Statistical significance level of 5%. ^†Statistical significance level of 10%.

Robustness Checks

The stronger results of samples in Models 3, 4, 7, and 8 may be driven by dropping observations that do not lend support for the hypothesis. Many countries in South America and East Asia democratized after the mid-1970s. As restricting the empirical analysis to years after 1974 runs the risk of letting Kuwait and Saudi Arabia drive the results, I drop both Kuwait and Saudi Arabia from the analysis and reevaluate the relationship between resource income and immigration policy openness. I use the binary regime classification in Models 9 through 12.

Table 8 shows results similar to those in Table 6. Compared with Model 1, the coefficient of resource income in Model 9 is more statistically significant. Dropping Kuwait and Saudi Arabia does not reduce empirical support for the hypothesis. The correlation between resource income and immigration policy openness remains robust with similar coefficients across Models 3, 4, 11, and 12 that test the hypothesis for years after 1974.

Table 8.

Authoritarian Immigration Policy Regressed on Resource Income (Autocracy = 1) Without Kuwait and Saudi Arabia.

	Model 9	Model 10	Model 11	Model 12
Years included	1946-2013	1946-2013	1975-2013	1975-2013
Immigration policy_t−1	0.841***	0.810***	0.645***	0.612***
Immigration policy_t−1	(0.028)	(0.027)	(0.031)	(0.070)
ln (Resource income/citizen)	0.004*	0.002	0.010***	0.016***
ln (Resource income/citizen)	(0.001)	(0.002)	(0.002)	(0.003)
ln (Population)_t ₋₁	−0.228*	−0.280^†	−0.628***	−0.108*
ln (Population)_t ₋₁	(0.063)	(0.130)	(0.105)	(0.416)
ln (GDP per capita)	−0.040*	−0.050*	−0.126^†	−0.141^†
ln (GDP per capita)	(0.018)	(0.019)	(0.065)	(0.065)
GDP growth	−0.065	0.009	−0.063	−0.081
GDP growth	(0.082)	(0.045)	(0.138)	(0.192)
Polity score		0.001		−0.006
Polity score		(0.001)		(0.004)
Observations	354	308	196	160
Countries	11	10	11	10
R ²	.882	.885	.816	.834

***

Statistical significance level of 0.1%. **Statistical significance level of 1%. *Statistical significance level of 5%. ^†Statistical significance level of 10%.

Finally, I use an alternative measure of immigration policy openness to reevaluate the empirical validity of the hypothesis for autocracies after 1974. Instead of the factor score, I use a standardized average of 7 entry and labor-market criteria: nationality, skill, quota, recruitment, labor prohibitions, deportation, and enforcement. The correlation between the immigration policy factor score and the standardized average is .76 for the 13 autocracies from 1946 to 2013. I use both the binary indicator of autocracy and polity scores (Polity < 6) to pool autocracies into two samples.

Table 9 shows that the results remain robust to the use of an alternative measure of immigration policy openness and different coding rules for autocracy. These additional tests illustrate the robust relationship between autocrats’ immigration policy choices and the extent to which they can use natural resource rents to keep native citizens quiescent.

Table 9.

Authoritarian Immigration Policy (Standardized Average) Regressed on Resource Income.

Regime classification years included	Model 13	Model 14	Model 15	Model 16
	Polity < 6	Polity < 6	Autocracy = 1	Autocracy = 1
	1975-2013	1975-2013	1975-2013	1975-2013
Immigration policy average_t−1	0.831***	0.820***	0.803***	0.801***
Immigration policy average_t−1	(0.029)	(0.041)	(0.024)	(0.030)
ln (Resource income/citizen)	0.010**	0.010**	0.009***	0.008***
ln (Resource income/citizen)	(0.003)	(0.003)	(0.001)	(0.001)
ln (Population)_t ₋₁	−0.085	−0.005	−0.138	−0.108
ln (Population)_t ₋₁	(0.096)	(0.111)	(0.079)	(0.121)
ln (GDP per capita)	−0.007	0.015	−0.027	−0.024
ln (GDP per capita)	(0.036)	(0.036)	(0.031)	(0.033)
GDP growth	−0.129	−0.087	−0.186*	−0.199*
GDP growth	(0.162)	(0.186)	(0.071)	(0.076)
Polity score		0.000		−0.000
Polity score		(0.003)		(0.002)
Observations	232	196	268	231
Countries	12	11	13	12
R ²	.766	.764	.830	.831

***

Statistical significance level of 0.1%. **Statistical significance level of 1%. *Statistical significance level of 5%. ^†Statistical significance level of 10%.

Conclusion

This article was empirically and theoretically motivated by the wide variation in authoritarian immigration policy. The theoretical prediction and the rigorous empirical analyses with a new dataset on authoritarian immigration policy emphasize the importance of natural resource rents, as well as the effect of its subsequent distribution in shaping autocratic elites’ incentive and ability to bring foreign low-skilled workers. While the majority of migrant workers living in wealthy autocracies are low skilled, some immigrant workers pose direct labor-market competition to a subset of native citizens. In Saudi Arabia, immigrants tend to fill young natives’ positions, depriving them of social places to coordinate on political actions and to overcome collective action problems. Only after the Arab Spring and protests by young Saudis did the Saudi Arabian government begin to give in to popular pressure for more employment opportunities in the private sector by reducing its addiction to foreign labor, which in turn ironically led to a series of human rights violations toward migrant workers.

Another implication for government-sponsored injection of foreign labor into the economy is the role of working women and gender equality. While the literature notes the role of natural resource rents in diminishing employment opportunities for women (Ross, 2008), the mechanism through which resource rents lead to gender inequality is unclear. The distribution of resource rents may promote patriarchy by reducing the work incentive of native women. In the meantime, rentier governments rely on temporary immigration to meet the labor demand, depriving women of work opportunities. In contrast, the state monitors and controls male citizens by employing them under government branches. In rentier states, male citizens tend to have government jobs for which they are often underqualified (Crystal, 1990). Male citizens’ dependence on autocrats’ provision of welfare and employment opportunities increases their loyalty. In turn, rentier governments promote patriarchy to exercise control over individual households through male citizens’ loyalty to the regime.

In rentier economies, temporary migrant workers also provide inexpensive services to native citizens who demand more services and goods due to an increase in redistribution. Immigration reduces the labor cost of domestic services, construction, and retail businesses. In an economy with lower prices, autocrats’ real income from resource revenues is higher. In other words, autocrats can provide a smaller paycheck to each citizen when the overall price level of a bundle of goods and services is lower.

The cases of oil-rich autocracies in the Persian Gulf illustrate how natural resource wealth encourages and allows autocrats to rely on low-skill immigration, which in turn increases the chance of autocratic survival through their reliance on resource rents and temporary immigration. The political consequences of mass immigration in resource-rich autocracies pose a challenge to our understanding of why some autocracies persist while others fade into history. In the new age of migration, future research will focus on the mechanisms through which temporary migration can consolidate autocratic survival.

Footnotes

Acknowledgements

I thank the editors and anonymous reviewers for their insightful comments. I also thank William Clark, Robert Franzese, Desha Girod, Trevor Johnston, James Morrow, Margaret Peters, and Nicholas Weller for their comments and suggestions. My special thanks go to Marc Helbling and Ines Michalowski for their invaluable input on several drafts of this article. In addition, I acknowledge helpful feedback and insightful questions from the participants at the 2015 American Political Science Association (APSA) Middle East and North Africa (MENA) Workshop, “The Resource Curse in the Middle East and North Africa” in Doha, Qatar. Finally, I thank the research assistants who assembled the immigration policies of additional countries: Pauline Hilmy, Meredith Garry, and Jessica Soley. All errors remain my own.

Author’s Note

This article is dedicated to the memory of my father whose strong faith in the welfare-improving power of international migration led me where I am today. Grant No. DGE

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This material is based upon work supported by the National Science Foundation Graduate Student Research Fellowship under Grant No. DGE 0718128. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author and do not necessarily reflect the views of the National Science Foundation.

Notes

Author Biography

Adrian J. Shin is a PhD candidate in the Department of Political Science at the University of Michigan, Ann Arbor and a National Science Foundation Graduate Research Program Fellow.

References

Abella

M. I.

(1995). Asian labour migration: Past, present, and future. ASEAN Economic Bulletin, 12, 125-135.

Acemoglu

Robinson

J. A.

(2000). Democratization or repression? European Economic Review, 44, 683-693.

Acemoglu

Robinson

J. A.

(2001). A theory of political transitions. American Economic Review, 91, 938-963.

Acemoglu

Robinson

J. A.

(2006). Economic origins of dictatorship and democracy. New York, NY: Cambridge University Press.

Andersen

Ross

M. L.

(2013). The big oil change: A closer look at the Haber-Menaldo analysis. Comparative Political Studies, 47, 993-1021. doi:10.1177/0010414013488557

Barro

R. J.

(1999). Determinants of democracy. Journal of Political Economy, 107, S158-S183.

Bates

R. H.

Lien

D.-H. D.

(1985). A note on taxation, development, and representative government. Politics & Society, 14, 53-70.

Bearce

D. H.

Laks Hutnick

J. A.

(2011). Toward an alternative explanation for the resource curse: Natural resources, immigration, and democratization. Comparative Political Studies, 44, 689-718.

Beblawi

(1987). The rentier state in the Arab world. In Beblawi

Luciani

(Eds.), The rentier state: Nation, state and the integration of the Arab world (pp. 85-98). London, England: Croom Helm.

10.

Beck

Katz

J. N.

(1995). What to do (and not to do) with time-series cross-section data. American Political Science Review, 89, 634-647.

11.

Boix

(2003). Democracy and redistribution. New York, NY: Cambridge University Press.

12.

Breunig

Cao

Luedtke

(2012). Global migration and political regime type: A democratic disadvantage. British Journal of Political Science, 32, 825-854.

13.

Bueno de Mesquita

Smith

Siverson

R. M.

Morrow

J. D.

(2004). The logic of political survival. Cambridge, MA: The MIT Press.

14.

Cheibub

Gandhi

Vreeland

J. R.

(2010). Democracy and dictatorship revisited. Public Choice, 143, 67-101.

15.

Crystal

(1990). Oil and politics in the gulf: Rulers and merchants in Kuwait and Qatar. New York, NY: Cambridge University Press.

16.

Crystal

(1994). Authoritarianism and its adversaries in the Arab world. World Politics, 46, 262-289.

17.

Czaika

Varela

M. V.

(2014). Labour market activity, occupational change and length of stay in the gulf. Migration Studies, 3, 315-342. doi:10.1093/migration/mnu048

18.

Facchini

Mayda

A. M.

Mishra

(2008). Do interest groups affect U.S. immigration policy? (IMF Working Paper WP/08/244).

19.

Freeman

G. P.

(1995). Modes of immigration politics in liberal democratic states. International Migration Review, 29, 881-902.

20.

Givens

Luedtke

(2005). European immigration policies in comparative perspective: Issue salience, partisanship and immigrant rights. Comparative European Politics, 3, 1-22.

21.

Goldin

(1994). The political economy of immigration restriction in the United States, 1890-1921. In Goldin

Libecap

G. D.

(Eds.), The regulated economy: A historical approach to political economy (National Bureau of Economic Research Project Report, pp. 223-257). Chicago, IL: University of Chicago Press.

22.

Grossman

G. M.

Helpman

(1994). Protection for sale. American Economic Review, 84, 833-850.

23.

Grossman

G. M.

Helpman

(2001). Special interest politics. Cambridge, MA: The MIT Press.

24.

Haber

Menaldo

(2011). Do natural resources fuel authoritarianism? American Political Science Review, 105, 1-26.

25.

Haggard

Kaufman

R. R.

(2008). Development, democracy, and welfare states. Princeton, NJ: Princeton University Press.

26.

Hainmueller

Hopkins

(2014). Public attitudes toward immigration. Annual Review of Political Science, 17, 225-249.

27.

Hanson

H. G.

Spilimbergo

(1999). Illegal immigration, border enforcement, and relative wages: Evidence from apprehensions at the U.S.-Mexico border. The American Economic Review, 89, 73-92.

28.

Hatton

T. J.

(2010). The cliometrics of international migration: A survey. Journal of Economic Surveys, 24, 941-969.

29.

Hatton

T. J.

Williamson

J. G.

(2005). Global migration and the world economy. Cambridge, MA: The MIT Press.

30.

Hatton

T. J.

Williamson

J. G.

(2007). A dual policy paradox: Why have trade and immigration policies always differed in labor scarce economies? In Hatton

T. J.

O’Rourke

K. H.

Taylor

A. M.

(Eds.), The new comparative economic history: Essays in honor of Jeffrey G. Williamson (pp. 217-240). Cambridge, MA: MIT Press. 217-240.

31.

Jensen

Wantchekon

(2006). Resource wealth and political regimes in Africa. Comparative Political Studies, 37, 816-841.

32.

Johnston

(2015). Weapons of the marginalized: Authoritarian bargaining and economic sabotage (Working paper). Retrieved from https://docs.google.com/viewer?a=v&pid=sites&srcid=dW1pY2guZWR1fHRqb2huc3R8Z3g6MjFkZjM5MDU3MDdiYjVl

33.

Jones Luong

Weinthal

(2006). Rethinking the resource curse: Ownership structure, institutional capacity, and domestic constraints. Annual Review of Political Science, 9, 241-263.

34.

Jones Luong

Weinthal

(2010). Oil is not a curse: Ownership structure and institutions in soviet successor states. New York, NY: Cambridge University Press.

35.

Kobrin

(1980). Foreign enterprise and forced divestment in LDCs. International Organization, 34, 65-88.

36.

Leblang

Fitzgerald

Teets

(2014). Defying the law of gravity: Political economy of international migration. World Politics, 66, 406-445.

37.

(2007). Political development in Hong Kong: State, political society, and civil society. Hong Kong: Hong Kong University Press.

38.

Mahdavy

(1970). The patterns and problems of economic development in rentier states: The case of Iran. In Cook

M. A.

(Ed.), Studies in economic history of the Middle East (pp. 428-467). London, England: Oxford University Press.

39.

Marshall

M. G.

Gurr

T. R.

(2014). Polity IV project: Political regime characteristics and transitions, 1800-2013. Retrieved from http://www.systemicpeace.org/polity/polity4.htm

40.

Massey

D. S.

Arango

Hugo

Kouaouci

Pellegrio

Taylor

J. E.

(1993). Theories of international migration: A review and appraisal. Population and Development Review, 19, 431-466.

41.

Miller

M. K.

Peters

M. E.

(2015). Restraining the huddled masses: Migration policy and autocratic survival (Working paper). Retrieved form http://sites.google.com/site/mkmtwo/Miller,Peters-Migration.pdf

42.

Mirilovic

(2010). Explaining the politics of immigration: Dictatorship, development, and defense. Comparative Politics, 42, 273-292.

43.

Money

(1997). No vacancy: The political geography of immigration control in the advanced industrial countries. International Organization, 51, 685-720.

44.

Morrison

K. M.

(2009). Oil, nontax revenue and the redistributional foundations of regime stability. International Organization, 63, 107-138.

45.

Ortega

Peri

(2013). The effect of income and immigration policies on international migration. Migration Studies, 1, 1-28.

46.

Peters

M. E.

(2014). Trade, foreign direct investment and immigration policy making in the United States. International Organization, 68, 811-844.

47.

Peters

M. E.

(2015). Open trade, closed borders: Immigration policy in the era of globalization. World Politics, 67, 114-154.

48.

Przeworski

Alvarez

M. E.

Cheibub

J. A.

Limongi

(2000). Democracy and development: Political institutions and well-being in the world, 1950-1990. Cambridge, MA: Cambridge University Press.

49.

Ross

M. L.

(2001). Does oil hinder democracy? World Politics, 53, 325-361.

50.

Ross

M. L.

(2008). Oil, Islam, and women. American Political Science Review, 102, 107-123.

51.

Ross

M. L.

(2012). The oil curse. Princeton, NJ: Princeton University Press.

52.

Ruhs

(2013). The price of rights: Regulating international labor migration. Princeton, NJ: Princeton University Press.

53.

Smith

(2004). Oil wealth and regime survival in the developing world. American Journal of Political Science, 48, 232-246.

54.

Timmer

A. S.

Williamson

J. G.

(1998). Immigration policy prior to the 1930s: Labor markets, policy interactions, and globalization backlash. Population and Development Review, 24, 739-771.

55.

Wantchekon

(2002). Why do resource abundant countries have authoritarian government? African Finance and Economic Development, 5(2), 145-176.

56.

Wintrobe

(1998). The political economy of dictatorship. New York, NY: Cambridge University Press.

57.

World Bank. (2014). World development indicators. Washington, DC: Author.

58.

Yom

S. L.

(2011). Oil, coalitions, and regime durability: The origins and persistence of popular rentierism in Kuwait. Studies in Comparative International Development, 46, 217-241.

59.

Zolberg

A. R.

(1989). The next waves: Migration theory for a changing world. International Migration Review, 23, 403-430.