The effects of the IMF on expropriation of foreign firms

An exception is a working paper by Jensen et al. (2014), which we discuss later in the paper.

See Graham et al. (2014) who contend that host countries now engage less in expropriation and more in restricting the transfer and repatriation of hard currency or limiting the conversion of local currency by foreign investors in what they call transfer risk.

See also Eaton and Gersovitz (1984) who demonstrate that the threat of expropriation adversely affects FDI.

Sectoral factors may also influence the likelihood for nationalization such that foreign firms involved in often changing technology that possess high mobility are arguably less prone to expropriation than firms that use slower changing technology and possess limits in their mobility. However, sectoral questions, albeit very important, go beyond the topic of this paper.

See, for example, Jensen (2006, 2008), Li (2006), and Staats and Biglaiser (2012).

See also Elkins et al. (2006) who argue that the spread of BITs is driven by international competition among potential host countries for FDI.

The Overseas Private Investment Corporation – a U.S. government agency – is another external actor used to explain domestic compliance with property rights (see Biglaiser and DeRouen 2007).

For a history of the IMF, see James (1996) and Stone (2002).

That is not to suggest that the IMF has completely forgotten about the developed world, as shown by the recent economic crisis in Southern Europe. Because of the financial troubles especially in the developing world over the past three decades, however, most of the IMF’s attention is focused on developing countries.

For insights on why developing countries accept IMF conditions, see Vreeland (2003a, 12–16).

See also Dreher (2003, 101) who notes that non-compliance with IMF agreements is the norm among borrowers.

See also Dreher et al. (2009) who argue that Security Council membership lessens the number of conditions in IMF programs, thus making compliance easier.

Although expropriation arguably provides borrowers with more resources for servicing debts – an important consideration as IMF member countries seek loan repayment – we contend that the Fund’s members have an overarching interest in promoting economic and financial stability as well as international trade and investment, which nationalization likely threatens in the long run.

See Graham et al. (2014) and Li (2009).

Jensen et al. (2014) principally investigates how economic shocks shape leader decisions to expropriate and secondarily considers how access to IMF funding is a lever used to punish for nationalization. Our main IMF findings are consistent with Jensen et al. (2014) except that we use four different measures of IMF influence whereas Jensen et al. (2014) employ only one.

See also Simmons (2000) who finds that market forces, and not the IMF directly, are responsible for countries complying with commitments to respect protection of property rights for the sake of upholding their reputation internationally. We also investigated if the expropriated firm’s nationality (particularly if it was a U.S. firm) might influence the Fund. The expectation is that because the U.S. holds veto power on important decisions, the Fund might for political reasons punish nations that expropriated U.S. firms. However, we have too few observations to make such a determination because nationality data are only available from 1993 to 2006. For more insights on the politics of expropriation, see Johnston (2012).

See also Gould (2003) who inquires as to whether the Fund serves the interests of developing countries or defends the goals of its financial backers. See also McKeown (2009) who contends that the U.S. is normally the most influential member of international organizations.

Kobrin’s (1984) data are from 1960 to 1979; Minor’s data include the 1980–1992 period. Many thanks to Christopher Hajzler and Quan Li for sharing their data.

The Hajzler dataset covers expropriations that have garnered the most public attention. For smaller expropriations that may not receive coverage in the data, see, for example, data maintained by the United States Foreign Claims Settlement Commission, where US nationals can file claims against foreign governments (http://​www.​justice.​gov/​fcsc/​completed-programs).

IMF loan data are available at: http://​www.​imf.​org/​external/​np/​tre/​tad/​extarr1.​cfm. Following Biglaiser and DeRouen (2011), we count undrawn IMF loans issued between 1961 and 1979 as if the country did not borrow from the Fund because, unlike nearly all loans since the 1980s, the borrowers did not draw down any of the monies and the IMF lacked influence over borrowers because the loans did not appear to carry any conditions. As one IMF historian (Boughton 2001, 2006) points out, borrowers often used the pre-1980 loans as a means to gain the Fund’s seal of approval. Similarly, Babb (2006) maintains that these loans were a “symbolic gesture of good intentions” made by borrower countries that were not suffering from balance of payments crises. Including the undrawn loans, the results for IMF remains negative and statistically significant but IMF SBA is no longer significant but that is expected as the SBA loans issued in the 1960s and 1970s bear no resemblance to loans issued since the 1980s. The results are available from the authors.

We also considered including an independent variable measuring compliance with IMF loans. However, doing so would entwine independent and dependent variables since expropriation, the dependent variable, is a form of noncompliance.

Following Haftel (2010), we also include in our U.S. BIT count, whether the country has signed a free trade agreement with the U.S., as such an agreement should potentially give the U.S. added leverage over host countries.

We also examined whether the likelihood of expropriation increases within 5 years of a default on foreign bond debts. Similar to Tomz and Wright (2010), we find that there is little correlation between default events and the likelihood of expropriation. We also tested an interaction term between IMF loan agreements and bond default and again the results are not statistically significant. The results are available from the authors.

In modeling the effects of an IMF agreement on expropriations, we recognize the possibility that IMF participation may not be randomly distributed (see Chwieroth 2005; Edwards 2005; Jensen 2004, 2006; and Vreeland 2003a), which potentially leads to a systematic bias. We considered using an endogenous switching Poisson model but such an approach relies on several assumptions that are unlikely to hold in our context. We recognize the potential limitations but we believe the models used in this paper control for the main explanatory factors in the literature and help expand our understanding of the influence of the IMF on domestic policy.