Risk Is Relative: Heterogeneous Responses to Institutional Risks for Foreign Investment

Are economic actors equally sensitive to institutional conditions? While existing research recognizes that institutions can have varying effects on actors’ interests, the implicit assumption is that actors are homogeneous in how sensitive they are to their institutional environment. We investigate this assumption in the context of foreign direct investment, arguing that actors from countries with weaker institutions will be less affected by information about host country institutional conditions—both good and bad. We test this argument using survey data from a diverse group of managers-in-training at an international business school. We ﬁnd that when asked to evaluate a potential foreign investment location, respondents from developing countries are signiﬁcantly less sensitive to information about the host country’s courts than their counterparts from developed economies. In contrast, we ﬁnd that economic actors from both developed and developing countries respond similarly to information about the stability of economic policies. The ﬁndings suggest that sensitivity to the risks and safeguards of certain institutional conditions vary systematically across actors, depending on both the home environment to which economic actors have been exposed and the type of host institution. to information about the quality of political institutions than individuals from developed countries.


D A N I E L J. BLAKE IE Business School, IE University, Spain
Are economic actors equally sensitive to institutional conditions? While existing research recognizes that institutions can have varying effects on actors' interests, the implicit assumption is that actors are homogeneous in how sensitive they are to their institutional environment. We investigate this assumption in the context of foreign direct investment, arguing that actors from countries with weaker institutions will be less affected by information about host country institutional conditions-both good and bad. We test this argument using survey data from a diverse group of managers-in-training at an international business school. We find that when asked to evaluate a potential foreign investment location, respondents from developing countries are significantly less sensitive to information about the host country's courts than their counterparts from developed economies. In contrast, we find that economic actors from both developed and developing countries respond similarly to information about the stability of economic policies. The findings suggest that sensitivity to the risks and safeguards of certain institutional conditions vary systematically across actors, depending on both the home environment to which economic actors have been exposed and the type of host institution.
Les acteurs économiques ont-ils tous la même sensibilité aux conditions institutionnelles ? Bien que les recherches existantes reconnaissent que les institutions puissent avoir divers effets sur les intérêts de ces acteurs, l'hypothése implicite est que la sensibilité de ces acteurs á leur environnement institutionnel est homogéne. Nous étudions cette hypothése dans le contexte des IDE, et nous soutenons que les acteurs des pays dont les institutions sont plus faibles seront moins affectés par les informations sur les conditions institutionnelles des pays hô tes, qu'elles soient bonnes ou mauvaises. Nous avons vérifié cet argument en nous appuyant sur les données d'une enquête menée auprés d'un groupe diversifié de managers en formation dans une école de commerce internationale. Nous avons constaté que lorsqu'il leur était demandé d'évaluer un lieu d'investissement étranger potentiel, les personnes interrogées provenant de pays en développement étaient considérablement moins sensibles aux informations sur les tribunaux du pays hô te que leurs homologues issus d'économies développées. À l'inverse, nous avons découvert que les acteurs économiques, qu'ils proviennent de pays développés ou en développement, réagissaient d'une maniére similaire aux informations sur la stabilité des politiques économiques. Nos conclusions suggérent que la sensibilité aux risques et aux mesures de sauvegarde de certaines conditions institutionnelles varie systématiquement chez les différents acteurs économiques, á la fois selon l'environnement d'origine auquel ils ont été exposés et selon le type d'institution hô te. economic outcomes. IPE scholars have built research communities examining the effects of preferential trade agreements (PTAs), bilateral investment treaties (BITs), domestic political constraints, regime type, and more (Pandya 2016;Baccini 2019). Such research emphasizes institutions' ability to shape the behavior of economic actors (e.g., firms, investors, and producers) by presenting them with different sets of opportunities, risks, or costs. In offering these institutional explanations, most extant IPE research implicitly assumes that these economic actors are both highly and uniformly responsive to the institutions they encounter. In this research note, we challenge that assumption. Using individual-level data, we explore how the importance of institutions varies across economic actors and ask: What makes some individuals' economic decision-making more or less sensitive to institutional conditions? Recent research has begun to consider that institutions can sometimes have differing economic effects. For example, political institutions may have greater risk-mitigating effects for multinational firms with higher fixed-cost investments (Kerner and Lawrence 2014). Similarly, institutions that advance trade liberalization may have varying effects on firms depending on their productivity (Kim and Osgood 2019). Emerging research shows that institutions may have differing effects if actors' backgrounds shape their abilities to perform better under some institutional conditions than others (Beazer and Blake 2018). We seek to complement such efforts, not by delving further into how institutions may affect actors' interests to differing extents, but by examining how actors vary in their sensitivity to institutional conditions. The difference is subtle but important. Sensitivity captures the degree to which institutions weigh on actors' decisions, and it is separate from whether those institutions protect or harm their interests. In this, economic actors' sensitivity to institutions cuts straight to questions of broad interest to IPE scholars: To what extent do institutions matter, and to whom do institutions matter the most?
We embed our study in the literature on institutional determinants of foreign direct investment (FDI). FDI scholars have long argued that, to protect their investments against political risk, investors favor host environments where political and legal institutions reliably enforce contracts, define and protect property rights, and promote policy stability (Nooruddin 2011;Staats and Biglaiser 2012;Li, Owen, and Mitchell 2018). In this FDI context, we ask whether and why investors might vary in their sensitivity to particular institutional settings.
To begin addressing this question, we investigate how investors' sensitivity to potential host countries' institutions is affected by where they come from. Recent research at the aggregate level demonstrates that home-country background conditions host-country institutions' effects on the location and quantity of FDI by shaping firms' relative skills and capabilities (Beazer and Blake 2018). Building on models of risk perception from risk analysis and management (Slovic 1987;Boholm 1998;Weber and Hsee 1998), we argue here that home-country experience can also significantly shape investors' responsiveness to host institutional conditions. Among international investors, the risks posed by "weak" host institutions-i.e., institutions that fail to provide predictable rules or effective and impartial enforcement-will seem more familiar and manageable to economic actors who themselves come from environments where "weak" institutions prevail. This suggests such investors will be less sensitive to institutional quality since, for them, the relative benefits and costs of strong/weak institutions are attenuated.
In this research note, we investigate heterogeneity in economic actors' sensitivity to host institutions using original, individual-level survey data from a diverse group of international economic actors: future managers-in-training at a top international MBA program. We focus on two sets of institutions that the literature identifies as significant determinants of investment: judicial institutions (e.g., Biglaiser and Staats 2010;Staats and Biglaiser 2012) and institutional constraints on policy change (e.g., Henisz 2000;Li and Resnick 2003;Jensen 2006). We find that survey respondents from developing countries-where institutions typically pose greater risks and challenges to firms-are significantly less sensitive to information about hosts' judicial institutions than their counterparts from developed economies. In hypothetical scenarios, providing negative statements about courts' effectiveness decreased investment recommendations among Organization for Economic Co-operation and Development (OECD) respondents by double that of their non-OECD peers. 1 In contrast, respondents from both developed and developing countries react similarly to information about the stability of economic policies. Our survey data suggest one explanation for these diverging results: Weak home institutions only reduce perceived risks when familiarity shows investors that the associated challenges are manageable. Although existing theories implicitly assume investors are equally responsive to institutions, our findings reveal that investors' sensitivity to host institutions can vary systematically, depending upon the institution and the home environment to which economic actors have been exposed.
These individual-level findings hold implications for theory-building in any IPE research area emphasizing institutions' role in shaping economic activity. Whereas much research assumes a smooth and uniform relationship between institutional conditions and economic actors' behavior, this research demonstrates systematic variation in economic actors' sensitivity to institutions. Evidence that institutions likely have a stronger impact on some economic actors' decisions and behavior than on others' opens up new avenues for research for multiple IPE research agendas. For example, research studying the impact of sweeping institutional reforms, such as those imposed by the IMF, could investigate whether institutional shifts have more or less profound effects depending upon the sensitivity of the actors responding to the institutional changes. Likewise, varying sensitivity to institutional conditions may influence the willingness of firms and industry groups to push for and support the development of international institutions (Kim and Osgood 2019).
Within FDI research, IPE scholars have begun studying how FDI's response to institutions is conditioned by traits that raise or lower investments' exposure to political risks, such as asset specificity (Kerner and Lawrence 2014), supply chain structure (Johns and Wellhausen 2016), or mode of investment (Pandya and Leblang 2017). Our research pursues a separate, yet complementary, avenue-investigating how home-country background shapes investors' sensitivity to the institutions that affect political risk. In doing so, our paper joins a handful of political science studies highlighting the importance of home-country factors to FDI activity (Leblang 2010;Kim et al. 2015;Wellhausen 2015;Beazer and Blake 2018). Our study also contributes empirically by investigating home-country effects using an original survey of international managers-in-training from eighty countries. In contrast to extant research that primarily employs firmor aggregated country-level data on FDI flows and stocks (e.g., Li and Resnick 2003;Jensen 2006;Holburn and Zelner 2010;Beazer and Blake 2018), our survey allows us to examine how individual-level assessments of host institutions vary by home environment.
Finally, despite growing investment and trade from developing and emerging markets, IPE research has done little to evaluate how these countries' differing political and economic environments shape their business actors' attitudes toward institutions they encounter abroad. As such, our research encourages scholars to consider how well existing research generalizes beyond the world's developed economies. As the group of globally competitive firms becomes increasingly diverse, IPE research will benefit from better understanding how differences in political and institutional backgrounds affect the individual-level decisionmaking that drives the global economy.

Institutions and Economic Actors' Behavior
A significant and ever-growing body of IPE research is concerned with how institutions-international and domesticimpact economic activity. Such research often seeks to explain how various institutions shape the opportunities and risks that influence economic decision-making. Within the trade literature, for instance, scholars investigate how the GATT/WTO and PTAs reshape firms' incentives by lowering barriers and stabilizing policies that affect trade, thereby influencing how much firms trade, with whom, and what products they export (e.g., Mansfield and Reinhardt 2008;Baccini, Pinto, and Weymouth 2017;Spilker et al. 2018).
Researchers studying FDI have been particularly interested in institutions' effects on international investment decisions. Scholars have noted the potential for PTAs and BITs to promote FDI by making government policy commitments to multinational firms more credible (e.g., Büthe and Milner 2008;Kerner and Lawrence 2014). Others emphasize domestic institutions' importance, arguing that investors favor host-country institutions that mitigate firms' exposure to political risks by establishing predictable rules and constraining others' opportunistic behavior. Accordingly, scholars have investigated FDI's relationship with various host institutions, generally finding that more investment goes to countries with institutions that protect and enforce property rights (Li and Resnick 2003;Staats and Biglaiser 2012;Jandhyala 2013;Moon 2015) or reduce volatility via greater constraints on policy-making (Henisz 2000;Nooruddin 2011). 2 Research on trade, FDI and other activities in IPE generally assumes that once salient factors are accounted for (e.g., differences in industry or competitiveness), firms respond to institutional conditions in a similar manner. This implies that economic actors are uniform in how they assess the risks and opportunities posed by the same sets of institutions. Yet, recent work examining firm and aggregate FDI location data suggests that this might not be accurate. Firms from different-developed versus developing-countries appear to find different types of institutions attractive as their prac-tices and strategies are more attuned to familiar institutional environments (Beazer and Blake 2018).
In this research note, we investigate whether the country of origin also affects how sensitive economic actors are to institutional conditions. This differs from prior work in that we seek to understand not which institutions are more or less attractive to actors from different countries, but rather how responsive actors from different countries are to positive and negative information about the quality of institutions. Focusing on international investment, we ask, when evaluating the same potential host institutions, do investors from developed and developing economies form similar perceptions? Are economic actors who deal with "risky" institutions at home more or less sensitive to host institutions when investing abroad? Understanding this sensitivity and its variance across actors is relevant not just to FDI scholars, but also to any IPE researcher interested in the connection between the institutional environment and economic outcomes.

Home Country Variance
Our interest in the country of origin as a source of variance in sensitivity to institutions has both theoretical and empirical motivations. Within international business research, existing theories highlight home countries' role in determining firms' economic outlook, capabilities, and business decisions (e.g., Kostova 1999;Cuervo-Cazurra 2006;Holburn and Zelner 2010). Likewise, we expect homecountry institutions could be similarly shaping economic actors' sensitivity to institutions abroad. We develop our theoretical argument more fully in the next section; for now, we briefly explain our empirical motivations for investigating home-country effects.
The increasing diversity of multinational firms has created a pressing need to better understand how home country may affect global economic actors' decisions and behavior. Cross-national data evince this growing diversity. Using multiple investment indicators, Figure 1 plots the percentage of global outward investment over time coming from developing countries. In the 1990s, developing countries contributed about 5 percent of annual global FDI flows, but by 2015 that proportion had quintupled to 25 percent. We observe similarly large increases from developing countries in outward FDI stock (3.5 ×) and cross-border merger and acquisition (M&A) purchases, both in terms of M&A quantities and purchase values (3.5 × and 2 ×, respectively). Simultaneously, this trend has resulted in greater heterogeneity among investors' home institutions. Figure 2 shows that average levels of judicial independence and political constraints among FDI-sending countries have fallen as investment from developing economies has risen. The implication is clear: The baseline profile of what FDI-sending countries look like is changing, with more global firms coming from places where domestic institutions provide less predictability and protection.
With these changes, IPE scholars have an opportunity to ask new questions and explore how well our existing theories and findings apply to economic actors outside those developed countries with strong institutions. To do this, we must understand how these newer entrants participate in the global economy, including how they respond to institutions abroad. 3 Here, we investigate one particular channel  Note: Data on judicial independence are from (Linzer and Staton 2015), and data on political constraints are from Henisz (2000). Both measures are continuous and bounded between 0 and 1; higher values represent more independence (constraints).
through which home-country institutions might influence economic actors' FDI decisions: by shaping their sensitivity to potential host institutions.

Home Institutions and Variance in Perception of Host Institutions
Why might economic actors from developing and emerging markets have a different sensitivity to host institutions than those from developed countries? Drawing on social science research in related fields, we suggest one plausible pathway: Home environments shape how actors perceive the political risks associated with weak institutions. Risk perceptions are distinct from risk "tolerance" or "preferences," which refer studies examining home country have tended to emphasize home-host dyadic connections: i.e., the role of diaspora networks (Leblang 2010;Graham 2014), shared language (Kim et al. 2015), and shared nationality with wronged foreign investors (Wellhausen 2015). Despite IPE's deep interest in institutions, the potential impact of investors' home institutions on investment has been underexplored in political science (see Beazer and Blake 2018, for an exception). to one's comfort with undertaking higher levels of risk. Instead, risk perception refers to the level of risk one associates with a certain state of the world (Weber and Hsee 1998). Research in management, public policy, and risk analysis all demonstrate that risk perception varies across individuals and that perceptions of the same risks can vary across countries or regions (Slovic 1987;Boholm 1998;Bontempo, Bottom, and Weber 1998;Weber and Hsee 1998). To date, this comparative research has not examined how actors from different countries perceive political and institutional risks.
While several factors can influence risk perception, researchers have identified attributes of the risk itself as a critical component (Boholm 1998). Early research finds that risk perceptions are a function of characteristics that can be grouped through factor analysis onto two dimensions: "dread risk" and "unknown risk" (Fischhoff et al. 1978;Slovic 1987). High levels of dread risk are associated with a perceived lack of control and highly costly-even catastrophic or fatal-outcomes. In contrast, high levels of unknown risk are associated with uncertainty and the perception that a particular choice has unobservable or unknown consequences. Thus, individuals perceive particularly high levels of risk when an option displays attributes associated with high levels of dread risk and unknown risk. Likewise, individuals perceive less risk when an option's attributes are not strongly associated with one or both of those underlying risk dimensions.
This model of risk perception provides theoretical foundations for expecting that economic actors from the developing world might react to host institutions differently from individuals from traditional FDI home countries. Regarding political risk, this research implies that individuals are more likely to associate weak political institutions with a higher level of risk when they see the risks posed by those institutions as new, unknown, or uncontrollable. Compared to their developed-country counterparts, economic actors from emerging markets are more likely to have witnessed or operated at home under a variety of less-than-perfect institutions (Cuervo-Cazurra 2006; Holburn and Zelner 2010). Actors from such countries need not have direct personal experience of the particular effects of weak institutions and potential mitigation strategies to have knowledge of them. Because they learn and draw lessons about their institutional environment from the experiences of others around them, economic actors can learn both experientially and vicariously (Baum, Li, and Usher 2000). As a result of this learning, individuals from these environments may be less likely to see similarly imperfect institutions abroad as posing new or unknown difficulties (i.e., lower unknown risk).
In the same vein, economic actors from developing countries are also likely to be more familiar with practices and strategies that firms can pursue to mitigate the effects of weak institutions (Beazer and Blake 2018). For example, firms can internalize more of their operations and transactions to minimize their dependence on institutions (Feinberg and Gupta 2009) or they can turn to corruption and bribery to resolve problems when operating within a weak institutional environment. Extant research finds evidence that investors from more corrupt countries are more willing to invest in countries where corruption is greater because they are more comfortable with corruption and are better prepared to engage in bribery (Cuervo-Cazurra 2006). More generally, having witnessed the realities of weak institutions at home, economic actors from developing countries will be more aware of the availability of second-best remedies that mitigate the effects of, reduce exposure to, or substitute for, weak institutions (Rodrik 2008;Dorobantu, Kaul, and Zelner 2017). Thus, economic actors from developing countries may be less inclined to perceive risks from weak political institutions as uncontrollable or generating costs that are unbearable (i.e., lower dread risk).
As actors from developing countries perceive lower levels of unknown risk and dread risk from weak political institutions, we argue they will take a less deterministic view of the effects of institutions on the quality of the business environment than their OECD-based peers, downweighting the risks from weak institutions and, by extension, the comparative reduction in risks from more reliable institutions. This reasoning leads to the following relative expectation: Hypothesis: When assessing the investment attractiveness of a host country, economic actors from developing and emerging markets will be less sensitive to information about the quality of political institutions than individuals from developed countries.
In the following sections, we use original survey data on economic actors from eighty home countries to test these ideas and gain empirical insight into how homecountry background shapes responsiveness to host institutions abroad.

Survey Description
If researchers wanted to investigate home-country effects on investors' perceptions, existing microlevel datasets would not help much. Cross-national surveys like the World Bank's Enterprise Surveys (WBES) show how business environments vary across countries, but do not ask respondents about foreign locations. Alternatively, in surveys that do study managers' attitudes toward foreign institutions (Biglaiser and Staats 2010;Johns and Wellhausen 2016), respondents come from a single country, often the United States. Such choices are appropriate for the original studies, but data without home-country variation cannot be repurposed to make inferences about home-country effects. Thus, to study home-country differences in how actors assess host institutions, researchers need individual-level assessments of host institutions and variation in respondents' home. Our original survey provides both.
Ideally, one might study investors' heterogeneous perceptions by surveying multinational firms from the complete spectrum of home countries, but global-scale firm surveys are prohibitively costly for most academic researchers. Instead, we use a second-best sample: international business managers-in-training at a top-tier MBA program. Specifically, we use a survey of 623 international MBA students at a leading business school in Europe. The survey was administered online between the fall of 2013 and spring 2015 as a take-at-home class exercise in separate sections of a required strategy course that took place midway through the MBA program. Average response rates exceeded 80 percent. Most respondents are male (75 percent) and have a business education background (>45 percent). The median respondent is 29 years old with 7 years of work experience following their undergraduate degree, many in big firms such as JP Morgan, Bayer, Hitachi, and HSBC. The main work experience for 80 percent is their home country. Respondents come from all major economic sectors.
Although not yet company directors, international MBA students are training to assess opportunities in diverse global locales and practice thinking through factors that affect business investments. MBA students cannot answer with a CEO's authority or expertise, but they have sufficient experience to hold informed opinions about factors that affect business ventures. Importantly, undergraduate and graduate students (including MBA students) have been widely used in studies of risk perceptions (e.g., Holtgrave and Weber 1993; Weber and Hsee 1998). Studies have found risk perceptions between students and professional analysts to be consistent within the same country, suggesting that differences in training affect risk perception patterns less than environmental factors (Bontempo, Bottom, and Weber 1998). Accordingly, we believe international MBA students are well suited to provide insights into how economic actors from different countries respond to political institutions' impact on investment.
A key feature of the data is that our survey respondents represent a globally diverse group of professionals. Although the MBA program is based in Europe, only 40 percent of respondents come from Western Europe, North America, or Japan. The remaining 60 percent of respondents come from non-OECD countries, primarily in Latin America, Eastern Europe, Asia, and the Middle East. Figure 3 shows the geographic coverage of survey respondents' home countries. 4 With respondents from more than twenty OECD countries and over fifty developing countries, our sample includes a far wider set of countries than any comparable study on investment attitudes that we know of in political science, economics, or management. Ultimately, this diversity allows us to investigate how responses to potential host institutions vary based on economic actors' home environment.

Differences in Experiences with Home Institutions
We begin by examining the argument's underlying premise that OECD and non-OECD respondents differ in their exposure to institutional risk at home. Although our theoretical argument is potentially applicable to a broad set of institutions, we focus on two of the most prominent institutions in the FDI literature: constraints on policy change by the executive and judicial institutions. While political constraints are arguably the most frequently cited institutional feature that affects FDI (e.g., Henisz 2000; Li and Resnick 2003;Jensen 2006), recent work shows that the quality of judicial institutions is also significant, and ranked as a top concern by Our survey data reveal a significant variation in respondents' familiarity with institutional risks at home. Figure 4 compares OECD and non-OECD respondents' answers regarding judicial institutions and constraints on policy change at home. 5 OECD respondents are vastly more likely to see their home courts as impartial and effective (65 percent vs. 25 percent) and report that courts encourage investment in their home country (49 percent vs. 14 percent). 6 Likewise, few OECD respondents think their courts discourage investment, yet roughly half non-OECD respondents do (17 percent vs. 49 percent). Interestingly, OECD respondents agree less often that economic policies in their home country are stable (37 percent vs. 45 percent); however, non-OECD respondents are more likely to believe unstable policy deters investors from their country (25 percent vs. 59 percent). Methodologically, these results provide a validity check, showing that our respondents have opinions about their home institutions that coincide with prevailing views. Substantively, these findings increase confidence that non-OECD respondents are more likely than their OECD counterparts to have been exposed to weak institutions at home.
Meanwhile, Figure 5 provides evidence regarding respondents' familiarity with the costs and management of institutional risks. The non-OECD respondents are more likely to report that they or their employers have incurred losses from problems with domestic courts (37 percent vs. 22 percent). 7 Non-OECD respondents are also more familiar with working around institutional problems. While 42 percent of non-OECD respondents report having successfully resolved problems with home courts, only 24 percent of OECD respondents say the same. Responses to policy stability problems reveal a similar pattern. Overall, these findings support the notion that non-OECD respondents have greater familiarity with the costs of problematic political institutions and how these problems can be managed. Interestingly, the prevalence of losses and resolved problems are dissimilar across our two example institutions, raising the possibility 5 Complete question wording for each panel appears in the Online Appendix. 6 All comparisons in Figure 4 are statistically different at p ≤ 0.05. 7 All comparisons in Figure 5 are statistically significant at p ≤ 0.05. that investors' beliefs about risks' manageability could vary by institution. We return to this question later in the empirical analyses.
Together, Figures 4 and 5 underscore the home institutional environment as a potential source of investor heterogeneity. Economic actors from non-OECD countries appear more likely to understand the realities of operating under weak institutions, including the potential costs and the options for resolving problems that arise. We turn now to investigating whether non-OECD actors' extra insight into the challenges of weak institutions leads them to be more or less sensitive to host institutions than their OECD peers.

Heterogeneous Responses to Hypothetical Investment Scenario
To investigate actors' sensitivity to institutional conditions, we use a survey experiment design. We analyze responses to a hypothetical scenario asking respondents to act as consultants evaluating potential investment locations for a fictional multinational corporation (QSQ Global). In the prompt, respondents randomly received either positive or negative information about whether the potential location's courts help businesses protect their interests and resolve disputes.
To examine whether home-country effects vary across types of host institution, respondents also received a second set of randomly assigned statements indicating whether frequent changes to taxes and regulatory standards were more or less likely. Both sets of statements were assigned independently using block randomization based on the OECD/non-OECD status of respondents' home country to ensure that respondents from both groups were exposed to all treatment conditions in roughly equal proportions. The OECD/non-OECD binary behind the block randomization scheme is justified by the previous section's evidence that OECD/non-OECD status is a good proxy for respondents' home experience with weak institutions. Nevertheless, we demonstrate in the robustness section that the results are very similar if we instead disaggregate the data using respondents' specific home country. While respondents only saw their randomly assigned prompt, we present both positive/negative conditions here side by side in square brackets:  Imagine that QSQ Global is considering investing in a developing country that has a population of 18 million people and is considered politically stable. GDP grew by 4.5% last year, a little more than the average rate in its region. According to observers, the country's court system [helps businesses / can make it difficult for businesses to] protect their interests and resolve legal disputes quickly. The country's current tax rates and regulatory standards for your client's industry are similar to those in competitor countries. These policies [have not changed much in the recent past / have changed much in the recent past; some changes have increased businesses' costs while others have reduced them]. Experts believe that the country's political system makes it [unlikely / likely] that there will be policy changes in the near future.
Following the prompt, respondents were asked to evaluate investment attractiveness by issuing a tentative recommendation to invest ("Explore investment opportunities in this country") or not ("Find an alternative location for investment"). Responses to this question form the dependent variable. Rejecting even this mild option to support investment signals that respondents have considerable aversion to the location description. We also note that inferences about our quantity of interest-differences in reaction to host institutions-rest upon the randomized assignment of institutional descriptions, not upon whether the threshold for supporting investment is set by a high or low bar.

Discussion of Results
Before reporting how OECD and non-OECD respondents assess these scenarios, we first pool all observations together. When the hypothetical host's courts help businesses defend their interests and resolve disputes, roughly 77 percent of respondents recommend pursuing investment opportunities in that location. Only 54 percent of respondents recommend investment when confronted with weak courts ( = 23, p < 0.01). Similarly, 74 percent of respondents favor investment when political institutions ensure stable policies, compared to 56 percent when institutions engender frequent policy changes ( = 18, p < 0.01). 8 Thus, pooled results reinforce conventional wisdom that effective courts and a stable policy environment have a positive impact on Figure 6. The effects of negative information about host institutions on willingness to invest, by respondents' home environment. Note: The left-hand panel plots the effects of moving from positive to negative information about host institutions on willingness to invest, conditional on home country and controlling for covariates. The figure's right-hand panel compares the difference in effects across groups. Bands represent 95 percent confidence intervals. Tables are available in the Online Appendix.
investor attitudes. Disaggregating the data, however, we see important nuances that conventional accounts overlook.
To analyze heterogeneous perceptions, we compare differences across OECD/non-OECD groups in the effects of receiving negative versus positive information about host institutions. Because we obviously cannot assign individuals' home country, we use linear regression (OLS) to adjust for potentially confounding factors that may be correlated with both respondents' OECD/non-OECD home status and their investment decisions: respondents' sex, age, educational background, years of work experience following college degree, and personal level of risk acceptance. 9 We include fixed effects for course section and industry of work experience to account for unobserved heterogeneity and focus on variation within MBA cohorts and industry. Figure 6 presents the corresponding results graphically, comparing the reaction among OECD and non-OECD respondents to information about host institutions. 10 For both courts and policy stability, Figure 6's left-hand panel plots the effect within groups of going from the positive to the negative institutional scenario on respondents' probability of recommending invest. The figure's right-hand panel compares the difference in those effects across OECD and non-OECD groups.
After separating respondents by home environment, we observe substantial differences in how much weak judicial institutions affect respondents' willingness to invest. Reports of weak courts decrease investment recommendations among OECD respondents by double that of their non-OECD peers (32 vs. 16 percentage points). This difference is substantively and statistically significant (p = 0.048). Roughly 82 percent of OECD respondents recommend investment when presented with effective and helpful courts, implying that difficult and unhelpful courts would drop endorsement among OECD respondents to an estimated 50 percent. In contrast, non-OECD respondents react to in-formation about courts in a more subdued manner. In the data, 74 percent endorse investment under the strong courts condition, but this drops to an estimated 58 percent endorsement under the weak courts condition. According to these data, positive and negative information about host courts does not impact non-OECD respondents' evaluations to nearly the same extent as it does for OECD respondents.
Thus far, the results yield two important insights. First, the systematically different responses between these groups to judicial institutions provides evidence that home environment can shape investors' perceptions of host institutions. Second, these results provide insight into whether familiarity with weak home institutions makes investors more or less sensitive to host institutions. Consistent with our our hypothesis, respondents from home countries with weaker courts (non-OECD respondents) are significantly less sensitive to host judicial institutions than those from home countries with strong courts (OECD respondents).
We now turn to the other institution of interest. Figure 6's top panel demonstrates that actors from developed economies place greater weight on host judicial institutions, but the bottom panel reveals a different pattern for policy stability. Information about the increased likelihood of frequently changing policies influences both OECD and non-OECD respondents in a similarly negative manner: The difference in treatment effects between OECD respondents (21 percentage points) and non-OECD respondents (13 percentage points) is statistically indistinguishable (p = 0.33).
Why do respondents from OECD and non-OECD countries have different sensitivity to host courts but not to unstable economic policies? We offer a tentative explanation consistent with the theoretical framework presented earlier: The manageability of risks varies across institutions. If some weak institutions create obstacles that firms can-with effort and know-how-accommodate or avoid, then familiarity with those institutions reduces investors' perceptions of dread risk and diminishes concern about those specific host institutions. In our study, judicial institutions seem to fit this scenario. However, other institutions may create problems that are very costly or highly difficult to manage. Economic actors familiar with such institutions at home would perceive higher levels of dread risk surrounding similar institutions abroad. If policy volatility fits this second scenario, then non-OECD respondents may have less uncertainty about host institutions that foster frequent policy changes, but greater concerns about the problems such institutions create. This would lead both respondent groups to display similar levels of sensitivity to institutions' capacity to promote policy stability. This explanation is broadly consistent with prior research that contrasts the ability of economic actors to manage risks associated with property rights institutions and contracting institutions (Acemoglu and Johnson 2005).
Our survey data do not enable extensive testing of this explanation for why we observe different home-country effects for different types of institutions, but they do provide some suggestive evidence for our interpretation. We direct readers back to Figure 5, which features respondents' experiences with home institutions. Focusing on the non-OECD subsets, two observations are particularly relevant. In the first panel, more non-OECD respondents have incurred losses from policy volatility than from weak courts (51 percent vs. 37 percent, p = 0.001), and yet, if anything, fewer non-OECD respondents in the second panel have resolved problems from policy instability than from courts (36 percent vs. 42 percent, p = 0.12). Slicing the data to compare within institutions, we see that, with courts, non-OECD respondents have incurred losses and fixed problems in roughly equal proportions (37 percent vs. 42 percent, p = 0.24), but for policy instability, far more respondents have incurred losses than have fixed the related problems (51 percent vs. 36 percent, p < 0.001). Compared to weak courts, policy instability appears more likely to be associated with business losses and less likely to offer workable mitigation strategies. These patterns are consistent with non-OECD investors' home experience making them less likely to view policy instability as an unknown threat, but also more likely to see it as posing risks that are particularly difficult to manage. Although merely suggestive, these findings invite further research into both how political risks' manageability varies across different institutions and how that manageability affects investors' assessments of, and sensitivity to, similar host institutions.

Additional Analyses
Our findings are robust to alternative specifications and estimation choices. We begin by exploring the findings' robustness to different measures of home institutions. In the analyses above, we compare OECD versus non-OECD respondents. One reasonable concern might be that this dichotomy is too crude and categorizes respondents' home institutions based on assumptions about the quality of home institutions rather than direct measures. Therefore, we divide respondents based on external measures of home countries' institutions and repeat the analyses. Specifically, we use the latent judicial independence (LJI) measure of judicial independence by Linzer and Staton (2015) to capture respondents' home judicial institutions and the POLCON III data from Henisz (2000) to measure home countries' constraints on policy change. For each measure, respondents with home countries in the sample distribution's top third are coded as having "strong" judicial independence (or political constraints) and receive a value of 1; otherwise, they are coded as "weaker" and receive a value of 0. Figure 7 replicates the main analysis using this alternative measure. As before, we observe no statistically significant differences across groups in reaction to information about policy volatility, but we find again that positive and negative information about host courts provokes a significantly stronger response among respondents who come from home countries with more autonomous and effective courts. These additional results are consistent with our earlier findings.
We also examine an alternative measure that relies on respondents' own subjective assessments of their home institutions. For this measure, we use respondents' opinion about how investment in their home country has been affected by the courts and the frequency of policy changes. For each institution, respondents who believe the institution has helped investment are coded with a value of 1; those who believe the institution has hurt investment or had no effect are coded as 0. The subsequent results in Figure 8 resemble the original findings. Respondents with helpful home courts are significantly more sensitive to positive and negative information about host courts than their counterparts from home environments with unhelpful courts. The consistent results in  Separate from measurement concerns, one might worry that MBA students from OECD and non-OECD economies differ in ways that could confound our inferences. For example, perhaps respondents from developing countries are more risk-acceptant, and this drives the heterogeneous responses to the investment scenarios. Figure 9 dismisses this concern. By multiple measures, the distribution of risk acceptance is nearly identical across the OECD and non-OECD groups. 11 To address concerns about imbalances across OECD/non-OECD groups beyond risk preferences, we reanalyze our main tests using entropy balancing-a 11 In addition, the main analyses control for individuals' level of risk acceptance.
technique that uses iterative algorithms to reweight our non-OECD observations such that both OECD and non-OECD groups have identical means and variances on all covariates observed in the data (Hainmueller 2012). Reweighting observations to impose balance across groups produces substantively similar results as before. These results appear in the Online Appendix.
Finally, because the survey took place over an extended period, perhaps respondents who participated early on might have influenced respondents in later waves in some way. Although we believe this to be very unlikely, we reestimate the analyses using only the first wave of respondents (Fall 2013) to ensure results do not reflect "contamination" of later respondents by previous sections. Again, we observe no meaningful changes from our main findings. Results for this test also appear in the Online Appendix.

Conclusion
Much IPE research explores the impact of domestic and international institutions on economic activities such as trade and investment. This study offers a unique insight into how economic actors may vary in their sensitivity to such institutions. Focusing on FDI and using an original survey, we find that respondents from non-OECD countries are less sensitive to information about host courts than their OECD counterparts. These results cast the literature in a new light. If existing theories primarily reflect Western business experiences, and existing analyses rely disproportionately on data from OECD countries and firms, current IPE research perhaps best explains only one specific type of economic actor-one who may differ meaningfully from the growing segment of other global economic actors. At least concerning courts and legal institutions, IPE scholars should consider how to adapt and extend existing theory to incorporate preferences and behaviors of developing-country actors that diverge from their developed country counterparts.
We also find that, although actors' sensitivity to information about judicial institutions varied across respondent groups, sensitivity toward policy stability did not. We have posited that prior business experience under some flawed institutions-ineffective courts-can provide perspective or risk-management skills that make economic actors more willing to accept similar conditions elsewhere, yet exposure to other institutional problems-unstable policy-may simply familiarize actors with how costly or intractable such problems are. For instance, we might expect investors from high-corruption home countries to respond less strongly to corruption in a potential host since for them bribe-paying could be a familiar response to institutional corruption that travels relatively easily across contexts (Cuervo-Cazurra 2006). In contrast, poor regulatory quality and dealing with a byzantine bureaucratic system probably requires more context-specific knowledge (and perhaps ties with specific government actors), making these risk-management strategies more difficult to apply abroad. Thus, like policy instability, familiarity with that kind of home environment may not make investors any less sensitive to similar host challenges. We hope that future research can test this argument on a broad range of political institutions.
Our findings also suggest specific new avenues for future research, starting with efforts to better understand what factors besides home environment affect how actors form risk perceptions. Research on debt markets shows that investors' perceptions of borrowing governments are influenced by their perceptions of "peer" countries' creditworthiness (Brooks, Cunha, and Mosley 2015). Further research could reveal whether investors perceive a host's institutional risks differently if they have experience with similar institutions in a peer country. Another fruitful research area would be to examine the symmetry of sensitivity to institutional conditions. While our findings show differences in overall sensitivity, we do not know if some types of actors are more sensitive to positive or negative information and whether this varies with past experience.
Finally, our research also has implications for how scholars think about development policy. Institutional reform is a cornerstone of development policy advocated by international development organizations and developed country governments, promising that painful reforms now will yield greater capital flows later. However, our findings suggest the pressure to improve weak institutions may depend on the investors that governments expect to attract. For example, investment from Chinese firms has increased dramatically across Africa over the past few decades (McKinsey & Co. 2017). Our study suggests that if less institutionally sensitive Chinese firms continue to be a dominant source of new FDI, governments in Africa will see less to gain from institutional development as the marginal effect of reform on Chinese investment flows will be modest. Indeed, depending on a country's potential investor profile, it may even tolerate a weakening of institutions without losing much in terms of capital flows.

Supplementary Information
Supplementary information is available at the International Studies Quarterly data archive.