Report

Corruption and Cross-Border Investment: Firm-Level Evidence

Beata K. Smarzynska and Shang-Jin Wei

Most developing countries and former centrally planned economies are eager to attract foreign direct investment (FDI). Hence, understanding the determinants of FDI is important in practice as well as in theory. Moreover, for many countries, a primary benefit of FDI is the inflow of technological knowhow of the foreign investor. As the technological content of a given FDI is closely related to the ownership mode of the investor (e.g., joint venture vs. sole ownership), it is also useful to understand the determinants of the ownership mode. In this paper, we study a particular determinant of FDI, namely host country corruption, that has received relatively less attention in the literature on FDI but is crucial in practice. While it is difficult to quantify precisely, casual empiricism would suggest that the cross-country variation in corruption level is probably as large as the variations in corporate tax rate or labor cost, two commonly emphasized determinants of FDI.

The issue of corruption has become a prominent item on the agenda of international institutions and national governments. The OECD Convention on Combating Bribery of Foreign Public Officials in International Business Transactions, which was signed in 1997 and went into effect in February 1999, criminalizes bribery of foreign officials by firms from member countries. Yet indices produced by organizations such as Transparency International suggest that corruption is still a widely spread phenomenon. In this paper, we examine the consequences of corruption on cross-border direct investment. More specifically, we look into two separate effects of corruption simultaneously: a possible reduction in the volume of foreign investment and a possible shift in the ownership structure.

The literature on FDI is too vast to be comprehensively referenced here (see Caves, 1982, Froot, 1993, Balasubramanyam et al. 1996). A subset of the literature uses firm-level data to examine the choice of entry mode (for example, Kogut and Singh, 1988; Blomström and Zejan, 1991; Asiedu and Esfahani, 1998; Smarzynska, 2000). A few papers have investigated the impact of corruption on FDI. This is a relatively new area of interest, with principal contributions from Hines (1995), Henisz (2000) and Wei (2000a and b). Hines (1995) was the first paper that reported a negative effect of corruption on foreign investment. His sample was, however, restricted to U.S. multinational firms. As Hines (1995) pointed out, because U.S. had been until recently the only major source country criminalizing bribery to foreign government officials, the effect of corruption on U.S. multinational firms may not be representative of the effect on the universe of foreign investors. Henisz (2000) was the first to study both the FDI market entry and ownership mode (e.g., joint venture vs. wholly owned firms). His sample was also restricted to U.S. multinational firms, and hence could also be non-representative of the universe of multinational firms. Furthermore, Henisz examined market entry and ownership mode separately rather than simultaneously. These two decisions could potentially be inter-related. In terms of statistical results, the estimated coefficients on corruption in Henisz?s paper were mostly not significantly different from zero or with a paradoxical sign in the sense that higher corruption appeared to be associated with more FDI. Wei (2000a and b) used a data set on FDI that went beyond U.S. multinationals, but the data were aggregated at a bilateral national level rather than at a firm level. As a consequence, it could not study ownership mode and entry decisions of the multinational firms.

We believe that it is time to revisit this important question by putting various ingredients together. We use a unique firm-level data set encompassing multinational firms from both the U.S. and other countries, which will allow us to examine whether host country corruption discourages investment by foreign firms for reasons beyond investors? fear of legal penalty in the home country. In fact, we will check explicitly whether U.S. investment behaved systematically differently from firms from other source countries. In this regard, the paper will examine issues that could not be examined in Hines (1995) and Henisz (2000).