When the government changes the rules after you invest

In 2008, after Venezuela announced the nationalisation of Cemex's subsidiary, government officials walked into the company's offices with the army (Moschieri et al., 2024, p. 641). Four years later the two sides settled out of court: the government paid USD 600m, half of what Cemex had claimed. In Argentina, Vivendi was awarded USD 105m against a claim of 317m, and the government did not pay until 2013 (Moschieri et al., 2024, p. 640). If you plan to build a plant abroad, which means burying money in something hard to move, two questions are worth asking first: what happens to your bargaining power once the money is in the ground, and does a local partner protect you? This post reads the obsolescing bargain alongside eight multinationals in disputes with Latin American governments, then offers a checklist for a mid-size Vietnamese investor.

What does the obsolescing bargain say, and which assets does it hit?

It says an investor's bargaining power is strongest before the money goes in and weakens once it is sunk. According to Moschieri and colleagues, Vernon (1971) treated the relationship between a multinational and a government as a negotiation. At first the government may offer tax rebates, subsidies and favourable rules, expecting capital, technology and jobs. After the costly investment is made and technology has spilled over to local competitors, the government reassesses (Vernon, 1971; Moschieri et al., 2024, p. 629).

It may change industry rules, alter or withdraw permits, or adjust taxes and tariffs, moving the benefits to someone else. The paper calls this creeping expropriation. A government may also seize assets outright. Even when compensation is offered, firms often feel it does not cover their losses (Moschieri et al., 2024, p. 630).

Not every industry is equally exposed. In the paper's literature review, investors in manufacturing and high technology tend to keep their bargaining power, whereas in resource extraction, large infrastructure and essential services such as telecoms, water, electricity and gas it erodes faster, because leaving is costlier (Moschieri et al., 2024, p. 630). The authors chose sectors where sunk assets make exit "equally arduous and costly" for every firm (Moschieri et al., 2024, p. 633); two of the eight are telecoms.

Eight multinationals in Latin America: who stayed, who left, who got what?

ICSID, the World Bank's centre for investment disputes, recorded 352 cases worldwide in 2001–2012. The authors narrowed these to 30 in Latin America, in sunk-asset industries, without internal causes such as fraud. They approached the 30 firms, and only eight gave access deep enough to study (Moschieri et al., 2024, p. 633). The table below condenses Tables II and III of the paper, in USD millions (Moschieri et al., 2024, pp. 639-646, 657). "Claimed" is what the firm asked for, usually more than it expected to receive (Moschieri et al., 2024, p. 667).

Two groups emerge. Telecom Italia, Iberdrola, Vivendi and, in its early phase, Endesa followed one path: narrow access to information, a late response and little local support once hostility was open. The other four raised the alarm early, gathered information widely and mobilised support at home and abroad (Moschieri et al., 2024, p. 637).

Why does relying on a local partner slow the warning?

Because the multinational hands the scanning and interpreting to its partner, and what comes back is late, possibly filtered and shaped by interests that differ from the multinational's own (Moschieri et al., 2024, pp. 650, 662). Moschieri and colleagues call the result the liability of insidership (Moschieri et al., 2024, p. 627).

An unnamed Telecom Italia director looked back: "We were myopic. We delegated and we should have intervened in first person." He added: "When there is no partnership, you have a better understanding of what happens." He also said his impression was that the local partner knew but did not say, though it was only an impression (Moschieri et al., 2024, p. 653). At Vivendi, a manager said the two other partners, Roggio and Dycasa, "didn't help. They had their own, different agendas" (Moschieri et al., 2024, pp. 656-657). The authors do not say every partner acts in bad faith: a local firm cannot relocate or turn to international arbitrators as a multinational can, so it may have to protect its own interests (Moschieri et al., 2024, pp. 662-663).

The firms running their own operations saw signals where a partner could not reach: managers at Cemex and Repsol noticed some contractors paying late and some large customers changing supplier (Moschieri et al., 2024, p. 651).

Endesa was the hybrid: it left dealings with the central government to its partner but supported municipalities and communities. In the authors' reading, that difference may explain why Endesa renegotiated its tariffs and stayed five more years (Moschieri et al., 2024, p. 649). They do not tell firms to avoid joint ventures (Moschieri et al., 2024, p. 666).

How solid is this result?

As solid as a qualitative study of eight cases can be. Only 8 of 30 firms agreed, and the authors suggest the sensitivity of the disputes may be why (Moschieri et al., 2024, p. 633). It is four against four, and the paper's six propositions come from case analysis (Moschieri et al., 2024, pp. 650-659).

The authors tested other explanations themselves. All the firms had positive assets, cash and earnings when the dispute began, and political-risk scores did not separate the two groups. But the proactive firms had, on average, more years of operating experience in Latin America (Moschieri et al., 2024, pp. 667-668). Experience could sit behind both the choice of embedding and the outcome.

The line between the groups is not "100 per cent owned" either: Cemex held 70 per cent, but the rest was spread among about 6,000 minority shareholders, with no local partner in between (Moschieri et al., 2024, p. 645). Nor did "proactive" mean keeping the assets: Cemex gave up ownership and Repsol sold out (Moschieri et al., 2024, pp. 658-659, 642). And the authors say the countries could be classed as state-led systems, and that different choices may work better elsewhere (Moschieri et al., 2024, p. 665).

What should a mid-size Vietnamese investor check before putting money abroad?

Check six things. This is my own inference from the study, not checked against data on Vietnamese firms. The example is wholly invented: a cassava-processing company in Tay Ninh, its investment and its partner are my own creation. It plans to put USD 12m into a starch plant in a neighbouring country, in a joint venture where a local company holds 40 per cent and handles licences and relations with the provincial authorities.

This suits you if most of your capital sits in fixed assets you own abroad. It does not suit you if you only export through an agent: the sources used here do not cover exporters.

Going deeper: what does each source cover, and what is unchecked?

The post rests on the full text of Moschieri, Ravasi and Huy (2024), Journal of Management Studies, volume 61, issue 2, pages 627-685, from the open-access PDF, including the tables and both appendices; the page numbers cited are the printed ones. For the paper's two model diagrams only the text extracted from the PDF is available. The post does not open the works the paper cites, so what it says about earlier research comes through the paper.

Vernon (1971) is a book the post does not use directly. The account of the theory here is Moschieri and colleagues' summary, plus the site's framework summary The obsolescing bargain between MNE and host state; that summary may leave out some of Vernon's conditions. The library's case page has only a short description, so the details of the eight firms come straight from the paper.

The paper has a few slips. Page 659 gives Repsol's claim as USD 10,500m and says it received half, while Table II prints 1,050m and 500m; this post follows page 659. Table II also says Endesa kept ownership of Edesur, while the text says it kept 48 per cent of the shares, so the post only says Endesa "stayed".