doi.org
Arbitrators as advisors: Evidence from changes in investment treaty design - The Review of International Organizations
The international investment regime is increasingly reliant on the expertise of lawyers. What impact does the rise of lawyers have on the design of international cooperation? We argue that lawyers can advise states to create international investment agreements (IIAs) that lower institutional barriers, making it easier for investors to access legal dispute resolution. Using an original dataset of 479 arbitration practitioners registered with the International Centre for Settlement of Investment Disputes from 1974 to 2023, along with a newly constructed dataset measuring investors’ access to legal dispute settlement, we find that states with a larger pool of arbitration practitioners are more likely to amend their treaties in a way that eases investors’ access to Investor-State Dispute Settlement (ISDS). We also find support for our theory through interviews with arbitration practitioners. Our findings can explain why states maintain investor-friendly ISDS clauses, even after suffering losses from previous ISDS disputes. The growing reliance on lawyers can reshape and professionalize international cooperation beyond the purview of states.
![Figure 1. Candidate support. Note. Figure represented with 95% confidence intervals, n = 1,047.
While all policy positions produced less favorable results for the Black women candidate than health care for all (control), findings presented below in Figure 1 suggest that the two racially-oriented positions correspond to the lowest levels of political favorability for Carolyn Johnson, with both racial treatment variables being statistically significant.12 Relative to the control condition, support for the Black woman candidate declined significantly when she endorsed a Black health policy (95% CI [−0.527, −0.017]). Support also declined when she endorsed a Black maternal health policy (90% CI [−0.478, 0.034]). This suggests support for H1c and marginal support for H1d.
Since the scores are above 5, a facile viewing of these responses may suggest tension with foundational theory,13 but I would underscore that, without an alternative candidate for the respondents to rate, we cannot make any claims regarding relative favorability of other hypothetical candidates (put differently, the respondents may very well be predisposed to rating candidates on the higher-end of the 1–7 scale).14 What we can assess is how support varies across issue frames, using health care for all, an agenda historically favored among women voters and one that is neither racialized nor gendered, as a reference category.
Among the experimental conditions, we can also see from Figure 1 that support for a Black women candidate is strongest amongst respondents exposed to the maternal health treatment. However, the estimated effect of the maternal health treatment was substantively small and statistically insignificant (95% CI [−0.279, 0.235]), indicating that support did not differ from the control condition. Therefore, H1a is not supported.](https://cdn.bsky.app/img/feed_thumbnail/plain/did:plc:ob5i5qtz52ds4c7d5sgk22mu/bafkreiaf3rlppbl53idyvnikcuxv4y4epny2umm52jfz3zede67rvbjbp4)
