Introduction

National interests and the pursuit of global climate, development, and biodiversity goals (e.g., Paris Climate Agreement, Sustainable Development Goals, Global Biodiversity Framework) have accelerated external investments in coastal and marine climate, sustainable development, and conservation projects in the Global South (hereafter “external ocean investments”). Since 1990, many blue economy and other ocean sectors have expanded 5-to-12 fold1. Coastal and marine areas in Global South countries are often targets for this acceleration, given that these areas are generally associated with rich biodiversity, high climate vulnerability, and substantial opportunities for advancing blue economy development2.

While well-designed and equitably implemented externally funded projects can significantly improve local well-being and ecosystem health, many well-intentioned external ocean investments can generate new inequities and reinforce pre-existing contextual inequity3,4. Contextual inequity encompasses the pre-existing uneven vulnerabilities and inequalities stemming from social, political, environmental, and other local-specific conditions and structures that influence the equity and effectiveness of projects (adapted from ref. 5).

For locations with high contextual inequity and substantial external investments, the stakes are high. In these locations (hereafter “high-stakes locations”) there are not only significant risks of exacerbating inequities, but also opportunities to reduce them. For instance, in Abadoze, Ghana, the convergence of climate adaptation, industrial fisheries, and other development projects exacerbated local poverty, health risks, and climate vulnerability within the marginalized, resource-dependent communities4. Investors did not consider how the construction of a thermal power plant and a sea wall in conjunction with foreign industrial fishing could impact water quality, increase the cost of living, reduce land availability, and drive further exploitation of climate-vulnerable fish species4. Similarly, in the Bay Islands, Honduras, increased externally-driven interests in expanding conservation efforts and eco-tourism led to widening economic inequalities through elite capture and increased marginalization of Indigenous Garifuna communities6. Conversely, through partnerships and prioritizing local leadership, conservation and development investors working in coastal Mozambique were better able to collectively address local vulnerability and inequity through climate-smart agriculture and conservation initiatives, capacity strengthening projects to improve women’s access to credit, and collaborative monitoring programs to assess project impacts7.

Despite the risks and opportunities of directing billions in external ocean investment to high-stakes locations, little is known about the distribution, magnitude, and equity of these investments in relation to the environmental, social, and political contexts in which they are implemented8,9. Contextual inequity is both complex and multidimensional and therefore difficult to assess and measure consistently. To this end, we developed a novel approach for measuring and mapping contextual inequity in coastal and marine locations in the Global South and examined the relative levels and equity of external ocean investment. With a more complete understanding of the levels and nature of contextual inequity and its interplay with external ocean investment globally, funders and implementers can make more informed decisions to mitigate risks and identify opportunities to advance equity and sustainability in coastal locations globally.

Contextual inequity

We developed a novel composite index of contextual inequity (see Methods) based on 14 local and national-level indicators distributed across three components: 1) local social-ecological vulnerability (climate vulnerability, economic deprivation, degraded ecosystems, marine economic dependency, and marine nutritional dependency); 2) within-country social inequality (gender, income, and life-expectancy inequalities); and 3) weak national governance (low voice and accountability, political instability, government ineffectiveness, poor regulatory quality, weak rule of law, and weak control of corruption) (Fig. 1 and Table S1).

Fig. 1: Contextual inequity composite index.
Fig. 1: Contextual inequity composite index.
Full size image

Conceptual figure of contextual inequity composite index (pink) comprising weak governance (6 indicators; blue), social inequality (3 indicators; purple), and social-ecological vulnerability (5 indicators; yellow).

Results

Contextual inequity varied greatly both within and between countries and regions (Table S2, Figs. 2, S1–S3). Weak governance and high social inequalities were generally the main drivers of contextual inequity (Fig. S3), except for Samoa, Barbados, Dominica, and Croatia, where social-ecological vulnerability played a relatively greater role. Contextual inequity was particularly high along the coastline of Africa (Figs. 2, and S1, S3), with countries such as Somalia, Guinea, and Comoros having some of the highest overall levels of contextual inequity globally (Fig. S3, Table S2). There was also exceptionally high within-country variability in some countries, such as Brazil, despite having relatively low inequity at the national level (Figs. 2, S2–3, Table S2).

Fig. 2: Global distribution of contextual inequity.
Fig. 2: Global distribution of contextual inequity.
Full size image

World map showing the geographic distribution of contextual inequity (A) and its three components: weak governance (B), social inequality (C), and social-ecological vulnerability (D), where 0 represents the minimum index value, and 1 represents the maximum index value.

External ocean investment

Our analysis of 35,440 coastal and marine external projects revealed growing, yet unequal, investments among countries. Between 2010 and 2021, the Organization for Economic Co-operation and Development (OECD) reported that funders committed over USD 32.6 billion in foreign development aid towards the ocean economy, with an increasing share directed towards climate adaptation, sustainable development, and conservation relative to other sectors (e.g., general infrastructural development) (Fig. S4). Of these sustainability-related projects, funders directed most external ocean investment towards sustainable water supply and sanitation (40%), environmental protection (21%), fisheries (13%), transport and storage (6%) and disaster preparedness (2.5%) (Fig. S5). Investment was concentrated in certain geographies, with more than half of the total investment directed towards ten countries located mostly in Asia (Fig. S6 and Tables S3 and S4). For instance, Indonesia received investments of over USD 2 billion (7.6% of total investment) over the study period, with an average annual growth rate of USD 74 million/year (Table S4).

High-stakes locations

We identified high disparities in the magnitude and distribution of external ocean investment and its overlap with contextual inequity globally (Fig. 3A, B). Where high external ocean investment coincides with high contextual inequity, the stakes for either reinforcing or ameliorating inequity are amplified. Of the locations where data were available (84 countries; USD 29.2 billion investment), we identified these high-stakes locations (i.e., investment and contextual inequity levels above median) in 23 countries (27% of the countries investigated) (Tables S5 and S6). Here, funders directed over USD 13.8 billion or 47% of global external ocean investment between 2010 and 2021 (top-right; Fig. 3B and Table S6). High-stakes conditions are present across most of the world regions, including Asia (e.g., Myanmar, Indonesia), Africa (e.g., Kenya, Tanzania), the Americas (e.g., Mexico) and Oceania (e.g., Papua New Guinea). (Table S5). When examining total investment per capita, the top 15 countries are mostly Small Island Developing States (SIDS) (e.g., Maldives; Fig. S7 and Table S4). Myanmar, Bangladesh, and Kenya are some of the countries with the highest risk and opportunities, with particulary high levels of external ocean investment and contextual inequity (top 25% for both) (Tables S2, S4, and S5). Sixteen countries (e.g., India, Turkey) had high investment (USD 12.4 billion; 42% of total global investment) but low contextual inequity (bottom-right Fig. 3B and Tables S5 and S6). Conversely, 21 countries (25% of the countries investigated) had low levels of investment (and investment growth) despite high contextual inequity (top left; Fig. 3B), together representing only 6% of the total investment (USD 1.6 billion) (Table S6). These were mostly African countries (14 countries) and SIDS (e.g., Fiji and Comoros) (Tables S5 and S6).

Fig. 3: Global distribution of external investment and contextual inequity.
Fig. 3: Global distribution of external investment and contextual inequity.
Full size image

Global distribution and trend of committed investment in externally-funded coastal and marine climate adaptation, development, and conservation projects (external ocean investment) relative to contextual inequity and presence of equity-related terms in project descriptions (n = 84 countries, USD 29.2 billion). A Global distribution of total, national-level committed investment relative to contextual inequity between 2010 and 2021. Color quadrants: green (high investment, low contextual inequity), light gray (low investment, low contextual inequity), pink (low investment, high contextual inequity), brown (high investment, high contextual inequity: “high-stakes” locations). Countries colored in ivory are countries that received investment during the period, but for which contextual inequity could not be estimated due to missing data (e.g., Egypt). B Level and trend of total, national-level committed investment in relation to contextual inequity. The size of the dots indicates the annual level of investment (USD million/year), and color represents the significance of investment trends (red: decreasing, blue: increasing, black: uncertain) (Table S4). The colors of the quadrants are the same scheme as in (A). White dots represent countries in the top 25 quartiles for both committed investment and contextual inequity. C Percentage of climate adaptation, development, and conservation (CADC) projects mentioning equity-related terms in project descriptions by country. The color of dots is the same scheme as in (A).

Equity in external ocean projects

Despite increasing attention to equity in global policy10, only 27% of the 35,440 reviewed projects explicitly mention equity-related terms in their project descriptions. Furthermore, equity considerations were lacking in more than half of projects in every country except Somalia (Fig. 3C). Encouragingly, mentions of equity are highest in high-stakes locations (e.g., Cambodia, Haiti, Tanzania), albeit not significantly greater than in other locations (Fig. S8). Since 2010, an increasing share of projects has made explicit mention of equity (Fig. S4). Nonetheless, equity mentions were extremely low in some low-investment/high-inequity countries, particularly in Africa (e.g., Gabon, Comoros) (<15%) (Fig. 3C). With expectations of increased development as Africa transitions from a carbon sink to a carbon source11, more equitable and effective external ocean investment is urgently needed.

Discussion

Strategic investments to advance ocean equity

In addition to confirming geographic disparities in ocean investment and contexts8,9, our findings highlight opportunities to strategically invest in improving both project and local contextual equity in various regions. Depending on their various capacities and goals, investors can prioritize projects or activities that reduce social inequality, social-ecological vulnerability, or improve governance in underfunded, high contextual inequity locations (top-left; Fig. 3B), particularly those where equity is scarcely mentioned (Fig. 3C). They may also consider coordinating investment strategies to promote a more equitable allocation of limited resources across countries. While external ocean investments are often driven by different political or social agendas, levels of risk tolerance, institutional structures, and constraints12, strengthening equity considerations and coordination should be central priorities for all investments in the Global South. However, advancing equity and improving coordination are even more critical for investors working in high-stakes locations (right and top-right, respectively; Fig. 3B) given the substantial levels of current investment and inequity in these areas.

For those interested in working in such heavy investment and high inequity contexts, investors can apply insights from this study on the levels and nature of investment and inequity to inform strategies and policies that mitigate negative impacts and advance equity. At the national level, weak governance and high social inequality in many high-stakes and/or high contextual inequity locations could indicate issues related to recognitional, procedural, and distributional equity. Weak governance, coupled with high levels of social inequality, may indicate a failure to recognize the rights, voices, worldviews, and impacts of policies on various socially and economically disadvantaged groups. For example, the combination of low voice and accountability and high gender inequalities reported in countries like Yemen and Sudan could suggest that women are being excluded from decision-making processes (procedural inequity) and that their rights, experiences, and perspectives are not being adequately recognized (recognitional inequity)13,14. This, coupled with high political instability, and extremely high levels of economic deprivation in many of these countries, may increase the vulnerability of women and other disadvantaged groups to further marginalization in poorly implemented coastal projects (distributional inequity). While it is encouraging that the mention of equity in project descriptions and levels of investment in coastal projects was higher in some of these locations relative to other countries (e.g., Somalia), in many other high-stakes locations, equity mentions are quite low (Fig. 3C). Thus, we recommend greater consideration of equity and coordination amongst the various investment actors operating in these countries. This could take the form of actively soliciting the input of marginalized voices (recognitional and procedural equity) to better tailor projects to meet their needs, avoid exacerbating preexisting inequities and vulnerabilities, and to facilitate more equitable sharing of project benefits and harms. (distributional equity).

Investing smarter and deeper to advance ocean equity in high-stakes locations

In addition to the risk of competing, counterproductive efforts in heavy investment locations, structural barriers such as power asymmetries, siloed governance, limited capacity, and lingering effects of past injustices often hinder efforts to improve sustainability and equity in coastal and marine policy9,15. To overcome such barriers in high-stakes locations, we propose that actors invest smarter—through diverse, well-designed and well-coordinated cross-sectoral partnerships, to better invest deeper—addressing the root causes of contextual inequity (Fig. 4). With this approach, and insights on the levels and nature of local contextual inequity, investors are better equipped to move beyond addressing ‘symptoms’ and progress towards the transformative change needed to reduce vulnerability and inequity and improve coordination in high-stakes locations.

Fig. 4: Potential approaches to invest smarter and deeper in high-stakes locations.
Fig. 4: Potential approaches to invest smarter and deeper in high-stakes locations.
Full size image

Examples of strategies for integrating contextual inequity considerations into coastal and marine projects to advance recognitional, procedural, and distributional equity in high-stakes contexts.

Through smarter partnership development, investors can collectively design cross-sectoral projects to invest deeper, addressing root causes that drive inequality, resource degradation, climate change, weak governance, and poverty in high-stakes locations. Smart and deep investments can help investors avoid reproducing the disjointed, externally-driven agendas that can lead to maladaptive outcomes and instead facilitate more strategic, equitable, and coordinated implementation of contextually-appropriate solutions (Fig. 4)16. For example, contextual inequity values may provide insight into structural factors affecting current efforts to support local organizations (e.g., weak governance), and investors can choose to divert funding priorities to address them (e.g., invest in peace-building vs. livelihood development). In other cases, investors may realize that some contextual inequity drivers are beyond their capacity, and may seek to partner with relevant investors for greater impact. For example, development actors can partner with conservation agencies in areas with high species vulnerability to support habitat restoration and recovery of socially or economically important species. In areas with high levels of social inequality, investors can leverage their collective influence to lobby for legislative change similar to the gender quotas that improved women’s representation in fisheries governance in Mauritius17.

Long-term investments in local capacity and leadership (e.g., female literacy and numeracy, higher education, and grass-roots organizations) could help strengthen the agency of marginalized groups across the multiple domains of climate, development, and conservation. Investors can also mitigate unintended harm and improve coordination through collectively developed social and environmental safeguards and monitoring and evaluation systems, which are often underfunded and disjointed across sectors. Knowledge and capacities generated in high-stakes locations through such partnerships could then be applied to areas that currently have high contextual inequity but low investment (top-left; Fig. 3B).

Beyond addressing local issues, investors can also focus on advancing systemic shifts within funding and implementing agencies and other governance systems at larger scales. This could involve multi-institutional efforts such as the Targeting Natural Resource Corruption project that aimed to address corruption, a root driver of both resource degradation and inequity12. Systemic shifts also include applying grant-making models centered on strengthening local tenure, rights, and agency (e.g., Turning Tides initiative18) and supporting or advocating for transitions amongst Global North actors away from policies and investments that undermine the well-being and resilience of vulnerable populations and ecosystems (e.g., unfair debt and trade policies, unsustainable fisheries subsidies, fossil fuel investment)19,20. By pooling their diverse areas of expertise and resources and working alongside local and global actors to identify priorities and potential leverage points, such partnerships are better positioned to advance equity across multiple scales.

Towards more equitable oceans

Just as healthy coastal and marine systems are essential to achieving climate, development, and conservation goals, so too is equitable external ocean investment. With billions of dollars being invested in coastal and marine areas in the Global South each year, failure to prioritize equity will not only undermine many economic and environmental goals but also the well-being of millions of people. Coastal areas are heavily used and thus highly political spaces, and advancing equity within high-stakes contexts is a formidable challenge. To address these challenges, we have proposed how to make investments smarter and deeper, ensuring they are successful in improving the well-being and resilience of people and the coastal and marine systems they depend on.

Methods

High-stakes definition

In this study, we identify “high-stakes” locations, areas with high (i.e., above median) levels of (i) coastal and marine climate adaptation, sustainable development, and conservation (CADC) projects (hereafter referred to as external ocean investment) and (ii) contextual inequity levels. To identify these locations, we investigated the context, magnitude, and attention to equity of external ocean investment within less economically developed coastal regions.

We conducted all analyses using R versions 4.2.2 and 4.3.2 software21. The R packages terra, sf, rnaturalearth, rnaturalearthdata, ggmap, rio, sp and rgeoboundaries are used for spatial data analysis and visualization22,23,24,25,26,27,28,29. EnvCpt was used to estimate investment trends30.

Data compilation

Coastal and marine climate adaptation, development, and conservation (CADC) investment

We used the database developed by the Organisation for Economic Co-operation and Development’s (OECD) Sustainable Ocean for All Initiative2 to examine the geographic distribution and equity of official investment towards coastal CADC projects. This database describes Official Development Assistance (ODA) for coastal and marine projects relevant to Sustainable Development Goal (SDG) 14 in and around 120 countries and territories within less economically developed regions31 (Table S4).

External investment categorization

The OECD Sustainable Ocean for All Initiative database also contains indicators for ODA directed towards the ocean economy and “sustainable ocean economy”, defined as coastal or marine economic activity that “emphasizes the sustainable use and conservation of natural capital in the world’s oceans, seas and coastal areas”2. This definition uses Sustainable Development Goal (SDG) 14 as a reference point, and as such includes projects that aim to increase or sustain marine ecosystem health and/or resilience through restoration, conservation, management, or sustainable economic use. This also includes land-based projects that seek to reduce land-based sources of pollution. Examples of “sustainable ocean economy” and “land-based economy” projects include sustainable fisheries management, marine protected areas, wastewater treatment, carbon sequestration, climate and disaster risk reduction, and marine research projects. Thus, focusing on OECD-classified “sustainable ocean economy” projects seemed appropriate to examine coastal investment at the climate-development-conservation nexus. We therefore used the “ocean economy” indicator to estimate total ocean economy investment and the “sustainable ocean” and “land-based” indicators to identify external ocean investment (i.e., CADC projects). After excluding bilateral projects (between two or more countries) where donor and recipient information were not available, the final dataset used for the study comprised 50,177 projects in 115 countries.

Investment trend assessment by country

To analyze trends in external investment by country, we employed a time series analysis approach using environmental change point detection32. We fitted a linear trend over time with temporal autocorrelation AR(2) errors on investment levels by country, allowing for the identification of potential structural changes in the time series. We extracted key parameters, including the trend coefficient and autoregressive terms (AR1 and AR2) from the model and calculated 90% confidence intervals for these parameters to assess the precision of our estimates. We then computed the adjusted R-squared value to evaluate the model’s explanatory power. We considered the temporal trend of investment significant when the 90% confidence interval did not cross the zero line, indicating either a positive or negative trend (Table S4). This methodology enables a robust analysis of external investment trends, accounting for temporal autocorrelation and providing statistical measures of confidence in our findings.

Identifying equity-related projects

In this study, we operationalized the concept of “ocean equity” grounded in the work of33. According to Blythe et al.33, ocean equity is defined as the recognition, meaningful involvement, and fair treatment of all relevant actors in ocean decision-making processes. It encompasses the fair distribution of marine harms, costs, and benefits, acknowledging the inherent right of all coastal people and communities to a healthy, productive, and sustainable marine environment. Three dimensions of equity—recognition, procedural, and distributional—are integral to this definition.

We identified projects that appeared to have equity considerations using the gender marker31 within the OECD database, and through a search string of equity-related keywords (Table S7). We developed and applied this search string using R to ascertain whether a project (based on its title and description) appeared to advance or address some aspect of recognition, procedural, and distributional equity. This includes projects that promote the rights and voices of various groups, seek to address social inequality, or improve the wellbeing of disadvantaged groups (e.g., poverty alleviation). We drew on the collective expertise of the researcher and practitioner co-authors to develop an initial list of equity terms. Subsequently, we expanded our keyword pool by incorporating terms related to SDGs 1 (No poverty), 5 (Gender equality), and 10 (Reduced inequalities) from research employing similar methods34,35,36.

While some keywords were consistently associated with only one equity dimension, others relate to multiple equity dimensions (Table S7). We used an iterative process to refine the search string to improve both its accuracy and precision. This involved assessing the returned results of each keyword when applied to the entire dataset or multiple random samples (25%). Keywords that successfully and consistently identified equity language on their own were assigned “high confidence”, those that identified equity language only when another equity term was included were assigned as “medium confidence”, and those that were inconsistently identified were assigned “low confidence”. We conducted additional tests to determine whether additional words could improve the precision and accuracy of “low confidence” terms, only including them if consistency improved. This iterative classification enhanced the robustness and accuracy of our keyword selection process and its alignment with the three equity dimensions: recognition, procedural, and distributional. The final list of keyword terms is detailed in the keyword list file (Table S7).

With the list of projects with equity considerations, we further categorized projects into three types:

  • “Equity CADC”: climate adaptation, development, and conservation (CADC) projects that explicitly mention equity-related topics (n = 9468 projects, USD 9.3 billion).

  • “Other CADC”: CADC projects that did not explicitly mention equity-related topics (n = 25,972 projects, USD 23.3 billion).

  • “Other ocean economy”: projects that were related to the ocean economy but not CADC projects (n = 14,737 projects, USD 13.7 billion).

A total of 35,440 projects (USD 32.6 billion) were classified as CADC (“Equity CADC” and “Other CADC”).

Contextual inequity composite indicator

Here, we defined contextual inequity as the surrounding social, economic, political, and ecological conditions and structures that shape people’s pre-existing status (e.g., wealth, social capital, assets/capabilities, and power) and, in doing so, subsequently influence the recognitional, procedural, and distributional equity in ocean projects (adapted from ref. 5).

Using this definition, we examined various local and national-level factors that are likely to influence equity within the local social-ecological context, and, as a result, influence project outcomes through their potential influence on, and sensitivity to, the equity of project design and implementation. We identified social-ecological vulnerability, social inequality, and weak governance as three general aspects of contextual equity that take into account inequities among countries (e.g., poverty rates, exposure to climate shocks), within countries (e.g., gender or income inequality, access to healthy marine ecosystems) and differences in the quality of governance structures and processes that are expected to shape equity (Fig. 1). We operationalized these three aspects through the 14 variables (see Table S1), creating scaled composite indicators for each aspect, and combined these into a single, scaled, composite index to represent local contextual inequity. Here, we explain how they strongly influence the recognition, procedural, and distributional equity of coastal projects.

  • Social-ecological vulnerability: externally-funded projects can support distributional equity by reducing social-ecological vulnerability within groups subject to marginalization. Distributional equity refers to the fair distribution of the benefits and mitigation of harms associated with the outcomes of ocean projects, actions, policies, and programs between different groups, including current and future generations. This includes efforts that promote the inherent right of all coastal people to a healthy marine environment that supports food security, livelihoods, culture, health, and well-being5,33,37. Areas where there is high nutritional and economic dependency on marine resources, along with severely degraded ecosystems, high deprivation, and/or climate vulnerability, would benefit greatly from effective and equitable external investment that seeks to improve ecosystem health, livelihoods, economic well-being, and climate resilience. Conversely, projects that limit access to resources (e.g., exclusionary, restrictive protected areas) or increase environmental harm or climate risk (e.g., removal of coastal habitats that protect shorelines), can exacerbate vulnerability, particularly amongst resource-dependent, disadvantaged groups.

  • Social inequality: The rights, values, and identities of women and girls living in areas with high gender inequality may not be adequately recognized or respected within governance institutions. Individuals with lower income and life expectancy may also be overlooked in coastal policy. Such social inequalities impact recognitional equity, which refers to the recognition and respect of human and customary rights, diverse cultural identities, knowledge systems, and values in ocean planning, management, and governance5,33,37. Poor recognitional equity may contribute to a failure to adequately account for the impacts of projects on women and other disadvantaged groups, leading to projects that exacerbate inequality and vulnerability. This is especially true if disadvantaged groups are also systematically excluded from decision-making processes, affecting procedural equity. Procedural equity refers to the inclusion and effective participation of all relevant actors in transparent and accountable decision-making for ocean actions, policies, and programs5,33,37. Conversely, projects that seek to address inequalities (e.g., national gender policies that require inclusion and/or consideration of women in decision-making processes) can lead to the inclusion and consideration of disadvantaged groups in decision-making, supporting recognitional and procedural equity, and are better positioned to improve distributional equity through more equitable design.

  • Weak Governance: Low voice and accountability in governance processes can contribute to disadvantaged groups being systematically excluded from decision-making processes, undermining procedural equity. Weak control over corruption or government ineffectiveness can lead to decision-making processes that are susceptible to political pressures by powerful actors (e.g., government officials, industry lobbies, or social elites), marginalizing others. Conversely, governance processes that actively support inclusion and engagement of disadvantaged groups in plural, transparent, and accountable decision-making related to external investments can lead to more equitable outcomes that advance social equity (e.g., climate projects that empower Indigenous Peoples and are influenced by Indigenous voices). Weaknesses in national governance, including poor regulatory quality, weak rule of law, and political instability, can also undermine the distributional equity of external investments by undermining efforts to support disadvantaged groups (e.g., improving the health of ecosystems used by resource-dependent groups) or by enabling elite capture of benefits. Conversely, measures that strengthen appropriate regulatory controls, fair rule of law, and political stability can help to ensure that benefits accrue to disadvantaged groups.

Our approach acknowledges this multifaceted and intersectional nature of equity, allowing for a more comprehensive understanding of the systemic contextual forces at play. We recognize that these 14 variables (Table S1) are insufficient to capture all aspects of vulnerability, weak governance, and social inequality, and that these three aspects do not fully reflect the full suite of conditions that contribute to contextual inequity. These indicators represent a comprehensive, but not exhaustive, set of metrics that are indicative of pre-existing location conditions that are influential over equity outcomes. The selection balances the need for broad coverage with practical considerations of data availability, spatial resolution, and representativeness across the study area. See the Caveats section in Supplementary Text for more details on data limitations.

Social-ecological vulnerability

  • Degraded ecosystems (Species at risk gravity)

    To create an indicator that measures the level of exposure of the coastal human population to coastal ecosystem degradation, we used the “species at risk” raster dataset that identifies coastal and marine species and ecosystems at risk from multiple anthropogenic stressors38. At the time of this study, this dataset represented the most up-to-date global assessment of cumulative human impacts on at-risk coastal and marine species for 1271 threatened and near-threatened species assessed by the International Union for Conservation of Nature (IUCN) Red List of Threatened Species38. We used this raster of the number of threatened marine and coastal species to generate mean gravity values for neighboring coastal (terrestrial) populated pixels within a radial buffer. This gravity concept captures the potential decreasing interactions of human populations with marine areas with increasing distance to the coast39.

    Given the computational resources needed to generate gravity values, we created a search buffer of 150 km to crop the species at risk raster around each coastal populated pixel. We then identified the nearest marine pixel by calculating a distance matrix from the populated pixel centroid. We then created a 20 km buffer radius around the marine pixel and estimated the mean number of threatened marine and coastal species affected by one or more stressors in the 20 km buffer. We chose a 20 km buffer as an estimate of the daily traveling distance of a small-scale fisher within a populated pixel to fishing grounds40. We choose small-scale fisheries given their importance for many human coastal communities41. We also hypothesized that most cultural and recreational ocean activities occur within a 20 km radius. The buffer size also accounted for computational constraints. For each coastal populated pixel, we estimated the species at risk gravity-adjusted mean by dividing the mean number of threatened species by the square of the distance to the identified marine pixel.

  • Coastal climate physical vulnerability

    To examine the vulnerability of coastal populations to climate change, we estimated the number of individuals living in low-lying areas. We sourced information on pixel-level population (5 km) from the Gridded Population of the World (GPWv4)42. We then constructed a coastal population raster at 5 km resolution within 100 km of the coastline. We chose the coastal population within 10 m of sea level as an estimate of coastal populations exposed to potential sea level rise.

  • Relative economic deprivation

    We sourced information on the relative economic deprivation from the Global Gridded Relative Deprivation Index, Version 1 (GRDIv1)42.

  • Economic and nutritional dependencies

We sourced information on the economic dependency on fisheries resources from ref. 43 and the nutritional dependency based on the per capita consumption of fish and seafood from ref. 44 at the national level.

Social inequality

  • Gender inequality

    We sourced the female’s share of total pre-tax labor income (2015), which considers gender differentials in earnings as well as labor force participation at the national level from ref. 45 as the gender inequality indicator.

  • Income inequality

    We sourced the income inequality from ref. 46 at the national level, averaging values from 2010 to 2022.

  • Life expectancy inequality

We sourced the life expectancy inequality index from ref. 46 at the national level, averaging annual values from 2010 to 2022.

Weak governance

We used the six World Bank Worldwide Governance Indicators (WGI) at the national level. These indicators combine data from survey respondents and experts from more than 30 sources, including households and firms, NGOs, commercial business information providers, and the public sector47. Each governance indicator was reversed so that higher indices reflect weaker governance. We renamed these indicators as follows: low voice and accountability, political instability, government ineffectiveness, poor regulatory quality, weak rule of law, and weak control of corruption.

Temporal data coverage

We chose 2015 (midpoint) as a benchmark for the majority of the contextual inequity indicators under the assumption of stability across the 2010-2020 period for those indicators. When we did not have 2015 values (4 of 14 indicators), we averaged across a concomitant period (2010-2020 or nearest available) to capture central tendencies where annual volatility was high. For the two variables with the greatest data constraints (Women’s Political Representation and Marine Employment Dependency), we used the nearest available investment period (2019) and the most relevant comprehensive dataset available (2003), respectively.

Operations on spatial rasters

We developed a national and pixel-level raster to combine the indicators used to create the contextual inequity index. We projected all 14 rasters using the Mollweide projection and transformed the national-level indicators into pixel-based data for raster analysis. We first resampled the pixel-level raster to a 5 km resolution, aligning with the extent and spatial resolution of the coastal climate physical vulnerability raster, using the bilinear resampling method. This method interpolates new pixel values by calculating a weighted average of the four nearest pixels in the original raster grid. The resulting value is based on a linear function of the distances between the target pixel and its neighboring pixels. Bilinear resampling produces smoother results compared to nearest-neighbor resampling, as it incorporates information from surrounding pixels rather than simply assigning the value of the closest one.

Next, we log-transform the individual indicators where necessary (e.g., species gravity, economic dependency, life expectancy inequality, and coastal climate physical vulnerability) to approximate a normal distribution. We then normalized each indicator using a min-max transformation to a [0, 1] range.

Computation of the composite index of contextual inequity

We used the 14 contextual inequity variables (Fig. 1; Table S1) to create scaled (0–1) composite indicators for social-ecological vulnerability, weak governance, and social inequality, and combined these into a single, scaled, composite index to represent local contextual inequity.

To create a comprehensive scaled, composite contextual inequity index at 5 km pixel that balances the multiple indicators and thematic groups, we employed a two-level hierarchical percentile ranking method48. This approach allows for a balanced representation of the various indicators while providing a single, interpretable index. Briefly, the method consists of first applying a percentile ranking function to each indicator (i.e., assigning rankings to each country for each indicator) within the three thematic groups: social-ecological vulnerability, weak governance, and social inequality. We then calculated mean composite indices and associated percentile rankings (second-level ranking) for each thematic group based on the first-level rankings (i.e., assigned new country rankings based on the average of individual indicator rankings). The final hierarchical composite index was computed as the mean of these second-level rankings. The final composite index ranged from 0 to 1, with higher values indicating higher levels of overall contextual inequity.

This index provides a comprehensive measure for comparing and analyzing contextual inequity across different coastal countries or regions. The dataset includes 115 coastal countries from various regions and economies for which data on the 14 indicators were available. This includes non-ODA recipient countries to provide a comprehensive range of both the lowest (e.g., Northern Europe countries) and highest possible values as reference points (Fig. 2).

Sensitivity tests

In this analysis, we chose equal weighting across indicators as an assumption-minimal approach for a general framework to avoid imposing subjective hierarchies that may not apply in all contexts. For illustrative purposes, we conducted two sensitivity tests to examine the stability of country rankings and classification with (1) the removal of specific indicators (i.e., leave-one-out analysis) and (2) randomly assigning weights to individual indicators.

  • Leave-one-out analysis

    A leave-one-out analysis tests for empirical influence and redundancy across indicators. When removing one variable at a time in the contextual inequity index, the marine resource dependency and social inequality indicators had the highest impacts on final contextual inequity scores, likely demonstrating redundancy between the governance indicators. We also observed large absolute changes in contextual equity composite scores for some individual countries when a variable was removed; overall, just under 52% of countries changed scores during iterations (Fig. S12 in Supplementary Materials). While these shifts appear substantial, no more than 12 countries shifted above or below the median value in each variable iteration, that is, indicating a shift to/from the top quadrant (e.g., to/from the top-right “high-stakes” category; Fig. 3B). Additionally, we only see minor changes in the overall average composite index score (<5%).

  • Random weighting

To further examine the sensitivity of the composite contextual inequity index to weighting, we carried out a Monte Carlo uncertainty analysis (100 draws), assigning random weights with a noise factor of 0.5 at both indicator and thematic group levels (e.g., social inequality) of the composite score (Fig. S13 in Supplementary Materials).

Gini coefficient

To examine variability in contextual inequity within each country, we calculated the Gini coefficient of contextual inequity at the country level using the DescTools R package49 (Figs. S2 and S3). A Gini coefficient is a measure of statistical dispersion. A coefficient of 0 indicates uniform contextual inequity across all pixels within a country, while a Gini coefficient close to 1 reflects considerable variability in contextual inequity within a country.

Final Dataset

As ODA data only applies to Global South countries, the final dataset included 84 countries (73% of countries in the contextual inequity database) where both contextual inequity composite indices and external investment data were available, covering 35,440 of 50,177 ocean projects (70% of total) and 29.7 USD billion (91% of total committed investment) in the CADC database.