Rights, institutions, and Chinese investment abroad

ABSTRACT

China has rapidly become one of the leading source countries for foreign direct investment (FDI), though many aspects of Chinese capital remain under examined. In this study we analyze the sociopolitical determinants of Chinese FDI, paying particular attention to the role of human rights and democracy. We hypothesize that Chinese capital is drawn to host countries with worse human rights conditions, given the lack of sensitivity that Chinese firms—particularly state-owned enterprises—have to potential “spotlighting” as well as the underdeveloped status of corporate social responsibility (CSR) practices. However, we posit that Chinese firms will be drawn to more democratic hosts due largely to experiential learning effects (negative experiences with autocratic hosts) and pragmatic advantages of investing in a state with democratic institutions. We test these hypotheses across 195 countries for the years 2005–2013, and find Chinese FDI to be negatively and significantly related to physical integrity rights and positively related to democratic institutions.

KEYWORDS:

• Foreign Direct Investment
• Human rights
• Democracy
• International political economy
• China

Highlights

• We assess the impact of host-country political institutions and human-rights conditions on Chinese investment abroad, including both private investment as well as investment by state-owned enterprises (SOEs).

• Examining Chinese investment flows from 2005-2013 we find that democracy has a positive and significant impact upon Chinese investment, particularly among host countries outside of the developed world.

• At the same time, we find that Chinese investment is drawn to hosts with lower levels of human rights (measures in terms of personal integrity rights).

Supplementary material

Supplemental data for this article can be accessed on the publisher’s website.

Notes

¹ Regarding labor laws, such efforts at developing a labor rights regime are somewhat apparent in extant data. For example, according to the Adams et al. (2016) data on labor rights regulations, China’s overall score for the years included in this analysis (17.74) falls within a standard deviation of the global mean value (19.83). By way of background, this index rates the presence of various regulations pertaining to labor rights and ranges from 0 (in which no rights are protected) to a possible maximum of 40. Yet China falls short in terms of actually protecting labor rights in practice; their score on the CIRI index of worker rights is 0 (the minimum score). Their record is uniformly low on personal integrity rights, and they score the minimum value (0) on the relevant CIRI index, and their average score on the human rights index used in our main analysis is over a standard deviation less than the mean.

² Democratic institutions are related to respect for human rights, but they are substantively and empirically distinct. Although human rights is ostensibly one of the things that a democratic government is supposed to protect, in practice the relationship between the two is sometimes more complex and qualified, as human rights repression can still occur even within established democracies (see Davenport, 2007). Broadly put, democratic institutions may hold states accountable for abusive behavior, yet governments may still be prone to repression directed towards groups that are less likely or able to impose political costs on states, such as ethnic minorities or otherwise disadvantaged groups. Indeed, some of the most abusive countries during the years covered here (as indicated by our human rights measures) were functioning democracies, including Colombia (with an average democracy score of 7 out of the range −10 to 10), India (9), and Israel (6). Moreover, each of these countries hosted major Chinese FDI projects across multiple years.

³ Regarding FDI into Africa it could also be the case that foreign aid could function as way to mitigate corporate risk, with the carrot of foreign aid potentially used as a tool to help mitigate risk (Busse et al., 2016; Dreher et al., 2018).

⁴ Given the project-centered nature of the data, as well as high cutoff for inclusion, missing data within panels was a problem. We thus used a fairly straightforward interpolation method in cases where data for a single country year was missing—in cases where a single year was missing we input the mean between the two surrounding years. If multiple years were missing for the first year we would interpolate between the previous value and 0. Substantively, interpolation is defensible given the path-dependent nature of FDI—that is, investors are more likely to continue to invest in countries with whom they have already established such relations. Moreover, studies of individual Chinese firms show that major projects seldom take place in isolation (see Liu & Li, 2002; Wang, 2014). Rather, as investors entering a new potential market, Chinese firms may first make smaller investments to gain a foothold in host country markets before proceeding with larger commitments. Also large investments may signal subsequent investments related to either further expansion into the market or simply to help provide support for the initial project. There is thus reason to expect that the $100 million cut-off may unduly truncate the actual investment patterns of Chinese firms. However, as a robustness check, we include parallel analyses of outbound FDI using individual projects as the dependent variable in our online appendix (see Table 2A). Results are largely consistent with those reported below.

⁵ Further descriptive information about patterns in Chinese FDI are contained in the Appendix (Figures F1A-F3A).

⁶ Descriptive statistics for this measure, as well as all variables used in the analysis, are in the Appendix (Table 1A). For more information about the construction of the human rights measure see Fariss (2014) and Schnakenberg and Fariss (2014). As a robustness check, we also run our models with an alternate measure of integrity rights, the CIRI measure. Results are unchanged from those reported below (see Appendix Table 5A).

⁷ In contrast to other measures of democracy, such as the Freedom House measures of political rights and civil liberties, the Polity measure is particularly useful as it focuses more clearly on democratic institutions rather than the practices associated with democratic governance. This enables us to more clearly differentiate between our democracy and human rights measures. Statistically, the correlation between democracy and physical integrity rights is .49.

⁸ The variable is constructed from the UN General Assembly Voting Data compiled by Voeten (2013); see also Bailey et al., 2017). Our measure is somewhat similar to that used by Duanmu and Urdinez (2018), who used average voting patterns between the host country and the US from 2005–2010, though our measure is not time-invariant and focuses more directly on Chinese interests beyond prospective balancing versus the US. In their study of Chinese FDI into Latin America, Tuman and Shirali (2015) use formal diplomatic recognition of Taiwan as a proxy. This was suitable given their emphasis on Latin America, only two countries outside of Latin America recognize Taiwan (Swaziland and Vatican City), neither of which is a host for Chinese investment. In sensitivity analyses we also ran models with an alternate measure, specifically the percentage of voting agreement with China. Results of these models were largely unchanged from the analysis presented below, as they were significant but in the opposite direction.

⁹ Specifically the lagged cumulative number of investment projects in a given country. In separate analyses, we ran models that included similar measures, fuel exports and ore and metal exports. Neither variable was statistically significant, and the inclusion of these variables did not alter the results of the main analysis.

¹⁰ To control for potential differences in investment patterns across regions we also include regional fixed effects (that is dummy variables for Western Europe, Eastern Europe, Latin America, Africa Asia). As there is a substantive reason to expect lagged investment levels to impact the amount of the location as well as the investment location, we also ran models with the lagged measure in the second stage. Results support the findings reported below. Including the measure in both stages was problematic for model identification purposes, as Heckman models “are appropriate only when at least one ‘extra’ explanatory factor influences selection but not the subsequent outcome” (Sartori, 2003, p. 112).

¹¹ As the Mill’s ratio in the fourth model did not rise to statistical significance, as a robustness test we ran this analysis with a simple two-step model, using probit regression for the selection and OLS regression for the amount phase. The results were similar except that the democracy measure no longer rose to statistical significance in the selection phase.

¹² Along those lines, while our study examines the determinants—rather than the impacts—of economic ties with China, extant studies do suggest some contrasting dynamics. Specifically, resource-intensive FDI may actually be helpful to human rights conditions in developing host states (Bader & Daxecker, 2015), though trade ties with China may undermine labor rights (Adolph et al., 2017). For our purposes, our analyses did not reveal any reciprocal relationship between Chinese FDI and either of our key independent variables.