2. Globalization of production processes and domestic property rights

Postwar economic globalization has promoted the international mobility of goods and production factors and cross-border integration of production processes. The fiscal and economic capability of national governments is increasingly intertwined with the globalized production and transactions of economic resources and commodities. A particularly salient aspect of economic globalization is the incorporation of foreign production factors in domestic production and value creation. Prominent foreign factors of production contributing to domestic value creation worldwide include fixed and financial capital, technology, and managerial skills and resources. While FDI and portfolio investment are essential vehicles of the international movement of capital, there are also other prominent circulations of non-capital components of production factors, such as international leasing of intellectual property rights and foreign labor compensation.

A key characteristic of foreign production factors is their international mobility, although some types of these factors may be more mobile than others. Taking capital factor as an example, factor inflows in the form of portfolio investment generally has greater mobility than FDI, which are more often, but not exclusively, associated with ownership of fixed assets in the host country (Jensen et al., Reference Jensen, Biglaiser, Li, Malesky, Pinto and Staats2012; Kerner and Lawrence, Reference Kerner and Lawrence2014). Production factors such as intellectual property rights and human capital are generally associated with even greater mobility. Because of the higher cross-border mobility, foreign factor owners acquire some bargaining leverage vis-à-vis the host government in demanding protection of their rights. The existing evidence on the domestic policy impact of FDI has highlighted international mobility and the credibility of the threat to exit essential factors in allowing the owners of the foreign factors to exert substantial influence on the host government regarding domestic governance quality (Malesky, Reference Malesky2009; Dang, Reference Dang2013; Demir, Reference Demir2016).

While foreign factor owners may derive preferential protection of property rights from their mobility, it is uncertain whether this protection would benefit other economic actors in the host economy more broadly (Wang, Reference Wang2015). Intense contributions of foreign factors of production in the domestic economy may enhance the overall quality of property rights protection because it would be difficult for the domestic authority to discriminate foreign interests from domestic interests in policy-making and implementation. When inward economic exposure is low, discriminatory treatment of economic actors by their national origins is administratively convenient and less costly. In such a case, the effect of foreign economic presence on the quality of property rights protection is likely to be limited to foreign economic actors. But with higher levels of integration of foreign production factors, it requires greater institutional capacity and resources to separate treatment of enterprises by domestic and foreign origins. When foreign enterprises receive better protections than their domestic peers, local private enterprises tend to strategically arrange their ownership structure to bring in foreign components to benefit from the preferential protection extended from the host government (Javorcik and Wei, Reference Javorcik and Wei2009).

The strong presence of foreign production factors provides distinctive opportunities for domestic economic actors to obtain better overall protection property rights. When foreign asset holders find legal and constitutional means of seeking protection more favorable than non-legal means, governments are more likely to establish universalistic protection of property rights rather than discriminating investors by their nationality and political connections (Wang, Reference Wang2015). In such a scenario, the property rights enhancing the effect of international mobility of production factors can extend far beyond foreign factor owners and benefit domestic economic actors more generally.

While the growth in the domestic presence of foreign factors may improve the domestic property rights regime as previously elaborated, an obstacle exists that may hinder their influence from being materialized. This obstacle is the lack of non-state actors' capacity for collective action under authoritarian rules. While the mobility of their assets and productive activities increases foreign economic actors' bargaining leverage, they nevertheless need collective action capability to induce the formation of a more general domestic property rights regime. Neo-institutionalist economists have long noticed that competitive market behavior alone may not be enough to constrain and shape the opportunistic behavior of powerful political actors (Greif, Reference Greif1993; Greif et al., Reference Greif, Milgrom and Weingast1994). Cross-border economic actors, suffering from poor information sharing and pervasive free-riding, are often incapable of organizing and collectively influencing governments to protect their interests (Greif, Reference Greif2006). In a more contemporary context, Wellhausen (Reference Wellhausen2013) has highlighted the lack of the alignment of preferences and coordinated action among direct investors and sovereign bondholders that renders their response to government predation ineffective.

The lack of collective action capacity among market actors is likely to be more pronounced in the authoritarian political context, where non-state economic actors have to bear high transaction costs to induce successful coordination and collaboration (Acemoglu et al., Reference Acemoglu, Verdier and Robinson2004; Bueno de Mesquita and Downs, Reference Bueno de Mesquita and Downs2005). The existing studies suggest that authoritarian leaders and elites have the incentive to co-opt and protect foreign economic actors who can, in turn, bring political and economic gains to the regime (Moon, Reference Moon2015; Bak and Moon, Reference Bak and Moon2016; Escribà-Folch, Reference Escribà-Folch2017). However, the capability of the economic actors to act collectively remains an essential condition that renders the commitment of the autocratic ruler to their protection credible. Taking in the insights from the literature on the origins of property rights institutions (North and Weingast, Reference North and Weingast1989), the credibility issue is essentially an incentive-compatibility problem: the government is better off in keeping its promise but may not be able to do so because of the enticing gains from defection. Institutions that empower the capital owners vis-à-vis the autocratic ruler make an essential hand-tying device to better align the government's commitment capacity with its incentives. Absent institutionalized constraints on the autocrat's power, economic actors with weak collective action capacity may not consider the authoritarian government's commitments credible as they have little power or influence to enforce them.

3. Institutionalized inter-governmental organizations and the property rights effect of production globalization

The institutional networks of IGOs remedy this collective action dilemma among non-state actors and strengthen the influence of foreign factors owners on property rights in authoritarian regimes. A few powerful IGOs such as the GATT/WTO, the World Bank Group, and the IMF have administered de jure policy constraints on member governments through the institutional procedures at the international level. Taking the World Trade Organization and the World Bank Group, for example, these two organizations have incorporated strong enforcement apparatuses to regulate the domestic practices of governments that pertain to property rights. In the WTO framework, arrangements such as the Agreement on Trade-related Investment Measures (TRIMS) and the Agreement on Trade-Related Aspects of Intellectual Property Rights (TRIPS) constitute binding regulations on government policies in the respective areas. More prominently, the International Centre for Settlement of Investment Dispute (ICSID) of the World Bank Group constitutes a cornerstone of the international investment treaty regime and investor-state dispute settlement (ISDS) mechanism, which extends significant influence and constraints on governmental practices in a wide array of policy areas. Memberships and participation in these powerful IGOs have been found to be able to shape the behavior of governments at the domestic level and contribute to subsequent improvement in credit and political risk ratings as well as macroeconomic performance (Tang and Wei, Reference Tang and Wei2009; Dreher and Voigt, Reference Dreher and Voigt2011; Dreher et al., Reference Dreher, Mikosch and Voigt2015). Institutionalized IGOs also impose de facto constraints that shape the policy behavior through a set of much more spontaneous mechanisms. In contrast to de jure constraints, de facto constraints are precipitated through foreign contributors' activities. Recent literature on global governance has seen a growing emphasis on the prominence of international market actors as the apparatus for enforcing institutional rules and commitments (Levy and Prakash, Reference Levy and Prakash2003; Kono, Reference Kono2007).

More specifically, institutionalized IGOs impose de facto constraints on authoritarian governments by establishing and maintaining various institutional arrangements, which enhanced the collective action capacity of international market actors when confronting authoritarian governments. Successful collective action among market actors hinges on the dissemination of critical information and effective coordination of initiatives countering the government's deleterious moves. Domestic actors and foreign factor owners operating in authoritarian regimes may have relative informational advantages to external actors and international organizations in assessing the policy, behavior, and intention of the government. But because authoritarian political institutions tend to forbid the formation of free and open information platforms, this information advantage of domestic insiders could not be transformed into an effective program for coordinating collective action among market actors more broadly.

Institutionalized IGOs could assist market actors in these respects through their permanent and institutionalized platforms for (1) information sharing and policy deliberations and (2) collective decision-making. A critical feature of institutionalized organizations resides in the functioning of a set of established procedures and rules for deliberative communication among their member states (Buchanan and Keohane, Reference Buchanan and Keohane2006; Smith and Brassett, Reference Smith and Brassett2008). These deliberative rules, oftentimes understood as critical for establishing legitimacy for institutionalization (Buchanan and Keohane, Reference Buchanan and Keohane2006), also embody a multilateral information regime that channels information from domestic insiders and disseminates them among transnational non-state actors (Nanz and Steffek, Reference Nanz and Steffek2004; Poast and Urpelainen, Reference Poast and Urpelainen2015). Such information and policy deliberation platforms facilitate the flows of otherwise private information on incentives and behavior of member governments, elevating the significance of state reputation (Guzman, Reference Guzman2008). The working of the deliberative function of institutionalized IGOs thus reduces the cost of information and helps non-state actors to respond effectively to predatory government practices (Gray, Reference Gray2009; Tallberg et al., Reference Tallberg, Dellmuth, Agné and Duit2018). Institutionalized IGOs provide an external platform for open deliberation and information sharing, which are otherwise not operable in the domestic context (Grigorescu, Reference Grigorescu2003; Poast and Urpelainen, Reference Poast and Urpelainen2015).

Furthermore, the rules and procedures for collective decision-making in institutionalized IGOs have the effect of internalizing the cost of collaboration among international market actors, mobilizing lobbying efforts, and elevating their bargaining power. The collective decision-making rules in IGOs devised to govern iterated negotiations within the organization lower the transaction cost for horse-trading and transfers among the members (Steinberg, Reference Steinberg2002; Davis, Reference Davis2004). The ensuing smoother interactions and transfers among the member governments help non-state stakeholders acquire much more coordinated incentives to lobby their home governments to pressure foreign counterparts for relevant policy concessions. Formal and institutionalized political procedures and rules have been shown to critically impact non-state actors' lobbying incentive and effectiveness (Naoi and Krauss, Reference Naoi and Krauss2009; Hojnacki et al., Reference Hojnacki, Kimball, Baumgartner, Berry and Leech2012). In the presence of the institutionalized collective decision rules, issue linkages are more easily established in inter-governmental concessions, which allow non-state actors to hold much more stable expectations about the lobbying outcome and coordinate actions more effectively (Gould, Reference Gould2003; Davis, Reference Davis2004). The collective decision-making institutions consequently enable transnational market actors to channel lobbying efforts efficiently and exert greater influences on the negotiations within institutionalized IGOs through their respective home governments.

A relevant example of the role of institutions and rules developed by IGOs that have helped foreign economic actors respond more effectively to predatory practices of host government comes from the system of international arrangements on anti-corruption measures. Facing corrupt practices from the government, foreign economic actors could collaborate to pressure the host government by publicizing and de-legitimizing such government practices. This strategy of response could be effective in pressuring the government (as they incur substantial reputation loss on the host government) but is rarely resorted to because of the daunting cost of collective action. Leaders and key coordinators and of these initiatives could be singled out by the authoritarian government and face harsh retaliations (Acemoglu et al., Reference Acemoglu, Verdier and Robinson2004). Since the early 1990s, several prominent international organizations, including the United Nations, the Organisation for Economic Co-operation and Development (OECD), and the Organization of American States (OAS), and the International Organization for Standardization (ISO), started to adopt conventions and standards that seek to standardize the treatment and prevention of government corruption among their member state. These conventions and standards include the UN Convention against Corruption, the Inter-American Convention against Corruption, the OECD Anti-Bribery Convention, and the Anti-Bribery Management System (ISO37001). Constructed on the basis of the pre-existing institutional infrastructures of international organizations, these anti-corruption arrangements formalize a global standard of anti-corruption rules and norms as well as establish a system of formal and institutionalized procedures for monitoring, review, and law enforcement with regard to anti-corruption initiatives. Under these arrangements, the political and operational costs of participation and contribution in the anti-corruption campaign are significantly reduced for foreign economic actors. Provisions of information for official reviews of host government behavior from foreign economic actors have been regular under the regime, precipitating a more unified stance among both domestic and foreign economic actors against predatory and corrupt government practices (Rose-Ackerman, Reference Rose-Ackerman2012).

The case of global anti-corruption arrangements across international organizations also demonstrates the significance of the policy networks formed among institutionalized IGOs in shaping domestic economic and regulatory policy (Cao, Reference Cao2009). The existing studies have highlighted several mechanisms of how policy diffuses between governments in international organizations' networks, namely socialization, emulation, and competition. The anti-corruption arrangements adopted across the networks of international organizations have seen these mechanisms of diffusion at work. First, anti-corruption conventions adopted across IGOs established norms that serve to socialize member states (McCoy, Reference McCoy2001). The institutions underlying the anti-corruption regime in monitoring and enforcement further allow foreign economic actors to provide information to generate normative pressure on governments to abide by the rules. This type of social and normative pressure produced by IGOs could have a substantial influence on the behavior of the government in countries sharing membership in the related IGO networks (Sandholtz and Gray, Reference Sandholtz and Gray2003). Similarly, emulation of responsible government behavior could also take place under these normative pressures. In the meantime, a competitive process among the member states could also take place in the global anti-corruption regime that facilitates the diffusion of good domestic governance practices. The regular publication of anti-corruption reviews and peer reports under the regime sends signals to prospective foreign investors regarding political risk and the business environment. Given that governments often compete with each other in attracting foreign investment inflows and the sensitivity of investment decisions to political risk, the anti-corruption regime across international organizations (Simmons and Elkins, Reference Simmons and Elkins2004) would prompt member governments to improve the business environment and pay greater respect to private property rights in order to attract more investment. Because the competition under the anti-corruption regime is centered on government practices in general, its positive effect on property rights will be materialized in the domestic context more widely. The network of IGOs under the anti-corruption regime fulfills both informational (Kinne, Reference Kinne2013) and behavioral functions (Alcacer and Ingram, Reference Alcacer and Ingram2013) that make up for the organizational disadvantages of foreign owners of production factor. Considering these propositions, we derived two hypotheses below to yield testable implications to be evaluated in the empirical analysis.

4. Foreign factors of production, IGOs, and property rights in authoritarian regimes: panel data evidence

Using a panel dataset from 1974 to 2005 at the country-year level, this part of the analysis seeks to empirically evaluate the association between globalization of production/inter-governmental organization and the protection of property rights in authoritarian regimes proposed in our argument. The time span of the sample is constrained by the data availability of key independent variables. Using Polity 2 score from the Polity IV project, we construct a sample of countries that are associated with a Polity 2 score <6, a common threshold used for demarcating democratic and non-democratic regimes between 1974 and 2005 (Cole and Marshall, Reference Cole and Marshall2014).Footnote ² We measure the quality of domestic property rights with CIM, which is a widely used cross-national measurement of the security of property rights developed by Clague et al. (Reference Clague, Keefer, Knack and Olson1999). As a financial indicator, CIM reflects the proportion of non-currency money that is held within banks. The logic underlying this measurement is that willingness of economic actors to deposit liquid assets such as money or quasi-money in financial institutions is ultimately tied to contract enforcement and the soundness of the property rights regime. The percentage of liquid assets that individuals choose to keep in banks is thus a reflection of the credibility of the government's credibility in protecting private property rights. Capturing economic agents' asset structures with easily accessible banking sector data, CIM comes with wide cross-national and time-series coverage that is not commonly seen in other existing measurements of property rights and the rule of law. Also, using standardized international banking statistics, CIM suffers less from the issue of subjectivity and the robustness of cross-country comparison than other property rights indexes. Another desirable feature of CIM is that it uses information coming from within the country, thus reflecting the degree of property rights protection that the domestic audience is under in general. Recent works that examine the cross-national variation in economic and property rights institutions using CIM include Fagerberg et al. (Reference Fagerberg, Srholec and Knell2007), Ahlquist and Prakash (Reference Ahlquist and Prakash2008), and Moon (Reference Moon2015). Table 1 lists a sample of CIM of authoritarian and democratic regimes in 2005.Footnote ³

Table 1. Property rights regime relative to political liberty in 2005

Source: International Monetary Fund, Polity IV.

Using CIM as the dependable variable, the model is specified with an auto-regressive model with lagged dependent variable formulated as follows.

(1)$$\eqalign{{\rm CI}{\rm M}_{i, \;t} = & \alpha \cdot {\rm CI}{\rm M}_{i, t-1} + \beta _1\cdot {\rm EXPOS}{\rm R}_{i, t-1} + \beta _2\cdot {\rm IG}{\rm O}_{i, t-1} + \beta _{12}\cdot ( {{\rm EXPOSR\\times IGO}} ) _{i, t-1} \cr & \quad + V_{i, t-1}^{\prime} \cdot \gamma + u_i + v_t + yr + \epsilon _{i, t}} $$

In this specification, ‘EXPOSR’ is the inward economic exposure which measures the integration of foreign production factors in the domestic economy. ‘IGO’ is the cumulative number of memberships in institutionalized IGOs with political and economic mandates. ‘IGO’, ‘EXPOSR’, and other control variables are lagged for one year. $V_{i, t-1}^{\rm ^{\prime}}$ is the vector of control variables. u _i and v _t each captures country and year fixed effect. yr is included to address potential biases stemming from a common dynamic trend in the dependent and independent variables. ε _i,t is the idiosyncratic error.

Several considerations motivate this specification. First, given the cross-country heterogeneity in economic structure and political institutions, endogeneity may underlie the association between property rights security and economic globalization due to unaccounted country-specific features. Similarly, endogeneity could also stem from systematic global forces that shape both property rights protection and the depth of globalization. Country and year fixed effects are therefore specified to absorb interferences from unobserved cross-country heterogeneity and global shocks. Under this specification, the results will be driven not by cross-country differences but within-country variations in the variables over time while controlling for time-specific systematic shocks, lending more validity to the finding.

The reversed causality between foreign economic exposure and CIM constitutes another concern. Methodologically, this reversed causal association will lead to a situation where foreign economic exposure is influenced by any accounted shocks to the quality of domestic property rights. Given these shocks are subsumed in the error term of the regression equation, our key independent variable, foreign economic exposure, will be correlated with the error term, causing biases to its estimation. We believe a combination of lagging the foreign economic exposure variable and controlling for domestic institutional and political covariates likely associated with property rights protection provides a feasible remedy to the problem. While lagging the independent variable seems simple, it should significantly weaken the correlation between the accounted shocks to property rights protection (at time t) and foreign economic exposure (at time t–1). After all, it is unlikely that shocks to property rights in a future period would affect the present level of foreign exposure. As we explicate when discussing the control variables, we included covariates capturing governmental features (e.g., new executive, and years in office), policy orientation (e.g., trade, investment, and financial openness, government size), and external assistance program (e.g., World Bank and IMF loans). Considering these covariates provide reasonable coverage of the observable sources of shocks to government policy and stance shaping domestic property rights environments, their inclusion may further attenuate the correlation between the accounted variation in CIM and the lagged observation of foreign economic exposure.Footnote ⁴

With the specified country and fixed effect, the auto-regressive model is substantively backed by the following characterization of the data generating process. The dynamics of the dependent variable, CIM, is conceptualized to follow a partial adjustment mechanism. That is, the actual response in the dependent variable, y _i,t, to changes in the independent variables is determined both by the equilibrium value of the dependent variable, $y_{i, t}^\ast$ (which is the ideal value of y _i,t if the adjustment is instantaneously accomplished), and the speed of adjustment, δ. This can be represented by the equation $y_{i, t}-y_{i, t} = \delta \cdot ( {y_{i, t}^\ast{-}y_{i, t-1}} )$. Incorporating the equation for $y_{i, t}^\ast{ = } \beta \cdot x_{i, t-1} + v_{i, t}$ and rearranging provides:

(2)$$y_{i, t} = ( {1-\delta } ) \cdot y_{i, t-1} + \delta \cdot \beta \cdot x_{i, t-1} + \delta \cdot v_{i, t}$$

The coefficient of the lagged dependent variable, 1 − δ ≡ α, can thus be interpreted as the reciprocal of the speed of adjustment. The partial adjustment mechanism is substantively relevant in the context of understanding the impact of institutional and economic factors on the private economic agents' evaluation of the security of their property. As the analysis relies on domestic economic agents' decision in structuring their assets for measuring the protection of property rights, it not only takes time for the economic agents to acquire information and update their belief of the strength of the government's commitment to private property rights but also takes time to restructure the profile of their asset as a result. The observed change in the dependent variable is likely to reflect an imperfect adjustment of monetary and financial assets within the time span of a year. Under this specification of the time structure, the idiosyncratic error is stationary.

4.1 Independent variables

Our main independent variable, ‘IGO’, counts the number of institutionalized IGOs of which the country is a current member. Construction of the variable uses the IGO membership dataset from the Correlates of War (COW) Project (Pevehouse et al., Reference Pevehouse, Nordstrom and Warnke2004) and codes relevant IGOs following the definition given below. An institutionalized IGO is defined as an IGO (1) whose main purpose or mandate is political, economic, or both, and (2) has incorporated deliberative collective decision and administrative organs that are governed by institutionalized rules and procedures. The focus on IGOs with political-economic mandates ensures the delegation to the organization is in relevant areas that are directly or indirectly associated to the interests of international market actors. Furthermore, such IGOs host regular deliberative activities with institutionalized arrangements such as assemblies consisting of permanent delegates from member governments or routinized (annual) ministerial meeting arrangements to precipitate the informational function of the organization. These IGOs have been governed by institutionalized organizational decision rules such as voting and or other formal procedures that provide a framework guiding intense political interactions among members. Finally, there should exist an independent bureaucracy within the organization to provide oversight of the fulfillment of the collective decision outcomes. Based on the population of IGOs provided in COW-IGO dataset, this selection criterion produced a total number of 119 institutionalized IGOs in 2005 as shown in Table 2.

Table 2. Total number and composition of IGOs: 1975–2005

The other key independent variable is ‘EXPOSR’, which measures the extensiveness of foreign contribution in the domestic economy or inward economic exposure. Conceptually, the extensiveness of foreign contribution in the domestic economy is most directly reflected by the portion of GDP yielded by factors of production owned by foreign individuals and corporations. While the existing GDP data do not allow distinguishing the nationality of the creators of the economic value, the inward economic exposure can be alternatively measured by factor returns earned by foreign factor owners using ‘primary income’ account data found in the balance of payments statistics of the IMF. Recorded in the current account, ‘primary income payment’ consists of compensations, dividends, interests, and rents paid to foreign persons and corporations. With the components displayed in Table 3, primary income represents ‘the return that accrues to institutional units for their contribution to the production process or for the provision of financial assets and renting natural resources to other institutional units’ (IMF, 2013). This measurement quantifies the contribution of foreign production factors (capital, human capital, and technologies) in the process of value creation and specifically the extent of the presence of foreign stakeholders in domestic economic governance. Regimes with the greatest factor income payment to foreign owners in 2005 are listed in Table 4.

Table 3. Primary income account

Source: International Monetary Fund.

Table 4. Domestic presence of foreign factors in 205

Source: IMF International Financial Statistics. Unit: billion USD.

The following control variables are included in the analysis. Politically, we control for the degree of domestic political constraints imposed on the government, utilizing the ‘executive constraints’ component variable from the Polity IV project. Economically, the size of the economy, per capita income, and growth rate is controlled. Additional indicators of economic openness, namely trade volume, and FDI inflows, are also included. Natural resource rent is included to account for the impact of resource abundance on economic structure and the pattern of governance.

While the country-fixed effect specification helps to mitigate endogeneity stemming from time-invariant country heterogeneity, endogeneity may still exist on other fronts. Most importantly, a within-country association between property rights and globalization may well be derivatives of more fundamental institutional changes or policy reforms occurring within countries. In the attempt to reduce this type of endogeneity, we selected a group of variables to be included to account for the evolving country-specific policy orientation and government characteristics. First, a dummy variable from the Database of Political Institutions (DPI) is included to indicate if a new executive came to power. This variable helps to capture the effect of a new government on policy reform and institutional transformation. Second, the consecutive years that the current executive has been in power are also included from DPI to capture the effect of the age of a government on policy agenda and orientation. Third, government size is measured by government spending as the percentage of GDP is collected from the World Bank's World Development Indicators (WDI) to capture the extensiveness of government intervention in the economy. Furthermore, the capital account freedom index developed by Chinn and Ito (Reference Chinn and Ito2002) is included to capture the general policy dynamics toward economic liberalization. We also included two variables, each indicating if a country was currently receiving any World Bank or IMF assistance. Finally, to control for the influence of the property rights-related agreements under the framework of WTO, we include a country's membership in GATT/WTO.

4.2 Results

The regression results are displayed in Table 5. Given the data generating process stipulated earlier, the coefficient of the lagged dependent variable provides the reciprocal of the speed of adjustment (i.e., α = 1–δ). Based on the results from Model (1) through (3), the average speed of adjustment is 0.304. Thus within the time span of one year, roughly one-third of the adjustment from the past to the new equilibrium level of CIM will be completed. As seen in Eq. 2, the true impact of the independent variables on the equilibrium level of CIM should be provided by the coefficients presented in Table 5 divided by the respective speed of adjustment.

Table 5. Foreign production factors, institutionalized IGO, and domestic property rights

Country-clustered standard errors in parentheses.

*P < 0.10, **P < 0.05, ***P < 0.01.

In Model (1), where no interaction term is included, ‘EXPOSR’ shows a positive but marginally significant association (at the 0.1 significance level) with CIM. This suggests a likely positive but weak average association between the contribution of foreign production factors in the domestic economy and domestic property rights security. The coefficient of ‘IGO’ is positive but insignificant in Model (1), suggesting membership in institutionalized IGOs alone carries no effect on CIM and domestic property rights protection more broadly. In Model (2) where an interaction term ‘EXPOSR × IGO’ is added, the coefficient of the interaction term gains a positive and significant sign at the 0.05 level of significance. This indicates the positive association between property rights and the domestic presence of foreign factors of production grows stronger when the government's membership in institutionalized IGOs increases.

In Model (1) and (2), constraints on the executive (‘EXCONST’) is shown to be positively and significantly associated with CIM, supporting the expectation that political institutions that effectively constrain the power of the authority is a sufficient condition for upholding a robust property rights regime. Among the control variables, per capita GDP and GDP growth rates are positively and significantly associated with domestic property rights at 0.05 level, consistent with the observation that developed economies are associated with a more robust regime of private property rights. In the meantime, the two alternative measurements of economic openness, namely trade and FDI inflow, are not shown to be directly correlated with CIM. The coefficient of resource rents gains a negative sign as would be expected but is not statistically significant. The size of the economy measured by total GDP is also not shown to be associated with CIM.

In the dynamic specification seen in Eq. 1 and 2, the marginal effect of ‘EXPOSR’ on CIM is provided by:

(3)$$\displaystyle{{\Delta {\rm CIM}} \over {\Delta {\rm EXPOSR}}} = \beta _1 + \beta _{12}\cdot {\rm IGO}$$

Using the output in Model (2) of Table 5, it becomes −0.00127 + 0.0000779 × EXPOSR. Figure 2 plots ΔCIM/ΔEXPOSR as a function of the cumulative number of IGO membership to further illustrate the conditional marginal effect. The marginal effect of EXPOSR is an increasing function of memberships in institutionalized IGOs. When the number of memberships in institutionalized IGOs is below 22, the marginal effect of the foreign contribution of production factors is statistically insignificant as the 95% confidence interval includes the horizontal line at zero. Growth in the integration of foreign production factors in the domestic economy is therefore not expected to be associated with enhanced levels of property rights protection when the government has low levels of participation in institutionalized international organizations. When the number of IGO members is >22, roughly the median of the sample, in contrast, the marginal effect of foreign production faction becomes positive and significant. Growth in the integration of foreign factors in the domestic economy shall be expected to associate with significant improvement in the domestic property rights regime.

Figure 2. Marginal effect of EXPOSR on contract-intensive money by IGO membership with 95% confidence interval.

Notably, the coefficient of ‘IGO’ is insignificant in Model (1), suggesting any unconditional effect of IGO membership on CIM is likely absent. Together with the significant moderating effect of IGO identified in Model (2), this finding potentially speaks to the different strengths of de jure and de facto constraints that the broader set of IGOs examined in this analysis impose on member governments. Because de jure constraints are enforced directed by institutional arrangements of IGOs that interact directly with member governments, their effect is likely to be manifested through a significant unconditional effect of ‘IGO.’ The insignificance of the unconditional effect of ‘IGO’ suggests the type of de jure constraints believed to be functioning under a few powerful IGOs such as the WTO, World Bank, and IMF is unlikely to be present for IGOs in general.

In Model (3) through (8) in Table 5, more political and policy-related variables are added to the specification to account for country-specific initiatives that may simultaneously associate with property rights policy and economic openness. These added variables are intended to capture the evolution of government characteristics which could potentially lead to changes in economic openness, IGO memberships, and protection of property rights. For example, when a new executive came to power, changes in economic policy and strategies for international involvement could well be more likely. If that is the case, the association between IPE-IGO and CIM may disappear after controlling for whether a power transition just took place. Model (3) includes a dummy variable indicating if a new executive just rose to power. The coefficient of ‘New Executive’ is nevertheless insignificant, while all the key results in Model (3) hold in Model (4). To address a similar concern, Model (4) includes a variable indicating the years the current executive stays in office. Because more durable leadership is more capable of enhancing the domestic regime for private property rights as the existing evidence suggested (Moon, Reference Moon2015), plus that the years in office is potentially associated with cumulative IGO membership, the correlation between IGO and property rights could potentially be spurious. ‘Years in Office’ in Model (4) indeed gains a positive coefficient that is significant at 0.05 level, indicating stable and durable leadership in an authoritarian regime is strongly associated with better protection of property rights. ‘Years in Office’ does seem to affect the relationship between IPE-IGO and domestic property rights, particularly the part conditioned by inward economic exposure, as the significance level of ‘IPE-IGO × EXPOSR’ changes from 0.05 to 0.10. The overall result from the models is nevertheless highly consistent with that from the original specifications.

Furthermore, Models (5) and (6) control for the size of government and capital account freedom which both directly reflect the policy orientation of the incumbent and could potentially shape international involvement and domestic property rights. Models (5) and (6) seek to rule out the possibility that the association between IGO and CIM is due to systematic liberalization of economic policy orchestrated at the domestic level. While ‘Government Size’ gains no statistical significance in Model (5), capital account freedom is shown in Model (6) to be positively and significantly associated with CIM. The statistical significance of the interaction term ‘IGO × EXPOSR’ somewhat weakens (significance level changes from 0.05 to 0.10) in Model (6), indicating part of the moderating effect may be associated with capital account liberalization. But all other key results in previous models continue to hold. Models (7), (8), and (9) further take into account the impact of participation in programs of the World Bank Group and International Monetary Fund and membership in GATT/WTO. Indicating if the government is part of any currently ongoing program administered by these two organizations, the two variables, ‘World Bank Program’ and ‘IMF Program’, intend to capture firstly the impact of the domestic crisis as well as the ensuing need for external resources, and secondly the potential push from participation in these programs for greater international involvement. Even though both are shown to have a positive and statistically significant (at the significance level of 0.05) association with property rights, the association between inward economic exposure, institutionalized IGO, and CIM is not weakened. Membership in GATT/WTO acquires a positive coefficient but no meaningful statistical significance. This result potentially suggests the effect of property rights-related restrictions and regulations imposed by agreements such as TRIMS and TRIPS is likely to be confined to areas that are closely related to international trade. Overall, the moderating effect of IGO membership on EXPOSR stays significant even after controlling for domestic and international level factors potentially associated with property rights reform initiatives and international outreach.

4.3 Robustness checks

We implement additional analyses to deepen the examination. First, we conduct a placebo-type test to examine the effect of institutionalization in IGOs. We do so by changing the dependent variable to memberships in IGOs of all types and levels of institutionalization. If the moderating effect of IGOs is indeed unique to those with high levels of institutionalization, the interaction term ‘IGO × EXPOSR’ should turn insignificant if we expand the IGOs to a broader set, including those with low levels of institutionalization. However, if the association between IGO memberships and property rights also holds for IGOs with low levels of institutionalization, the earlier results we obtain may be outcomes of spurious regression due to unobserved covariates that associate changes in property rights regime and a general pattern of participation in international organizations. This placebo-type analysis provides a chance for falsifying the aspect of our argument tied specifically to the institutional characteristics of IGOs.

Table 6 presents the results of the test. In Models (1)–(4), the IGO variable is the count of memberships in IGOs that do not meet the standard of institutionalization defined earlier. In Model (1) where no interaction term is included, the coefficient of ‘EXPOSURE’ continues to acquire a positive sign which is significant at 0.10 level. However, the previous finding on the moderating effect of IGO membership ceases to hold as the interaction term, ‘IGO × EXPOSR’, turns insignificant at 0.10 level in Models (2)–(4). In Models (5)–(8), where the IGO variable is the count of memberships in all IGOs with all levels of institutionalization, the outcome is similar: EXPOSR continues to gain a positive and weakly significant coefficient, while its interaction with IGO does not gain significance at 0.10 level. The other control variables generally gain statistical significance in a manner consistent with that in the previous results. This analysis lends support to the linkages between our theoretical claim and the empirical finding. While the beneficial effect of IGO memberships on property rights security exists for a reasonable range of IGOs with certain institutional features, it is, however, not extendable to a broader set of international organizations.

Table 6. Institutionalization in IGO: a placebo test

Country-clustered standard errors in parentheses.

*P < 0.10, **P < 0.05, ***P < 0.01.

We also employ alternative measurements to capture the quality of property rights protection using ‘the Rule of Law’ Index and ‘Property Rights’ Index from the ICRG. Quantifying the ‘strength and impartiality’ of the domestic legal system (Howell, Reference Howell2011), the Rule of Law Index has been used in the existing studies to provide measures of the rule of law as well as the legal capacity of political regimes. Property Rights Index is a composite measure combining the efficiency in government bureaucracy with the perception of corruption. The notable distinction between the two lies in that the former corresponds more to so-called ‘contracting institutions’ and the latter corresponds more to ‘property rights institutions’ in the framework provided by North (Reference North1981). In light of the work of Acemoglu and Johnson (Reference Acemoglu and Johnson2005), both contracting and property rights institutions contribute separately to the commitment capability of governments, which then shapes the ensuing developmental outcomes. As CIM constitutes a behavioral proxy for the outcome of domestic institutions, the two indexes from ICRG more directly capture different institutional and procedural aspects of the domestic commitment process.

The key results from the estimation as shown in Table 7 partly corroborate those from models of CIM. Most importantly, the moderating effect of IGO on EXPOSR is shown to exist in models of ‘Property Rights’, as the coefficient of the interaction term ‘IGO × EXPOSR’ is positive and significant at 0.10 level in Models (6) and (8) and at 0.05 level in Model (7). This result confirms the previous findings based on CIM. In contrast, the moderating effect disappears in the models of the Rule of Law Index as the interaction term ‘EXPOSR × IGO’ fails to acquire significance at 0.10 level in Models (1)–(4). While the differences in the outcome merit more sophisticated examination, this result likely suggests the effect of foreign economic exposure and global integration may have a stronger effect on property rights institutions, which primarily constrain the confiscatory incentives of the government, than on contracting institutions that reflect more of the legal capacity of the government. Based on the estimation results in Model (2), we calculated and plotted the marginal effect of ‘EXPOSR’ in Figure 3, where the pattern is generally similar to that in Figure 2. ‘EXPOSR’ has a positive marginal effect on Property Rights Index when the number of institutionalized IGO membership exceeds 29.

Figure 3. Marginal effect of EXPOSR on property rights index by IGO membership with 95% confidence interval.

Table 7. Alternative operationalization of property rights

Country-clustered standard errors in parentheses.

*P < 0.10, **P < 0.05, ***P < 0.01.