Climate change is causing substantial sea-level rise (SLR), posing significant risks to coastal communities globally. Effective adaptation strategies are crucial, and understanding the macroeconomic consequences of SLR and adaptation responses is essential for policymakers. Existing macroeconomic assessments of SLR often neglect or oversimplify adaptation, particularly the role of coastal migration. This study addresses this gap by providing a comprehensive analysis of the macroeconomic effects of adapting to SLR, focusing on the higher end of SLR projections until 2050. The research explores various adaptation options, including advance, protection, retreat (migration), accommodation, and ecosystem-based solutions, recognizing that the effectiveness of these options varies with the magnitude of SLR. High SLR may necessitate coastal retreat and out-migration, a crucial aspect often overlooked in macroeconomic assessments. This study aims to provide a more realistic assessment by incorporating both protection and migration as adaptation strategies and comparing the macroeconomic costs of these different adaptation pathways.

The literature on the macroeconomic effects of SLR has two primary limitations addressed in this paper. First, many studies assume a no-adaptation reference scenario, which is unrealistic since coastal societies have historically and will continue to adapt. Second, most studies focus on protection measures while neglecting the significant role of coastal migration as a retreat strategy. While some integrated assessment models (IAMs) and computable general equilibrium (CGE) models have included migration, they typically employ simplified representations that do not fully capture the complexities of migration dynamics and costs. Some IAMs use stylized national-level damage functions, while others model migration as forced consumption, ignoring productivity losses from moving capital. Existing studies highlight substantial costs of SLR-induced displacement, but the need for more comprehensive modeling is evident.

This study employs a soft-linked approach, combining the bottom-up coastal impact and adaptation model DIVA with the multi-sectoral and multi-regional global CGE model COIN-INT. DIVA assesses coastal flood damages, protection costs, and migration based on detailed data on population exposure, asset values, and extreme water levels. COIN-INT then incorporates these impacts into a macroeconomic framework, enabling analysis of economy-wide effects. Protection is modeled as planned, publicly financed adaptation, while out-migration is considered autonomous and privately driven. The analysis considers three scenarios until 2050 based on the RCP-SSP framework: RCP8.5-SSP5 with high-end and medium ice melting sensitivity, and RCP4.5-SSP2 as a middle-of-the-road scenario. Each scenario is analyzed under four adaptation cases: no adaptation, planned protection only, autonomous migration only, and combined protection and migration. The study compares these scenarios to SSP-specific baselines without climate change, isolating the impact of SLR under different adaptation assumptions. Migration is modeled as internal migration within countries, driven by flooding exceeding a 1-in-1-year event threshold. Costs include moving and rebuilding capital stock, while changes in spatial economic agglomeration are assumed to be neutral. The study uses GDP as an indicator of macroeconomic effects and examines direct and indirect effects, including capital accumulation dynamics and amplification ratios (ARs), which measure the relative size by which direct damages are absorbed or amplified through the economic system. Finally, the sensitivity of the results to ice melting assumptions is assessed.

The study's key findings demonstrate that adaptation clearly pays off economically, but residual damages can still be substantial. Under high-end SLR, combining protection and migration is more cost-effective than protection alone in regions like India and Southeast Asia. This is attributed to the high investment costs of sea dikes, which tie up capital that could be used more productively elsewhere. Autonomous migration alone substantially reduces macroeconomic losses, particularly in developing regions. Comparing planned adaptation to a no-adaptation scenario versus comparing it to an autonomous adaptation scenario reveals that the benefits of protection can be overestimated if autonomous adaptation is not considered. Direct costs (sea flood costs and migration costs) are significantly amplified over time due to capital stock dynamics and indirect effects. The amplification ratio (AR), reflecting the ratio of annual GDP losses to direct costs, often exceeds 1, indicating that indirect and intertemporal effects dominate the macroeconomic costs. In several regions, GDP losses under autonomous migration are less severe than under planned protection alone, highlighting the economic inefficiency of solely relying on costly protection measures. Sensitivity analysis shows that macroeconomic effects are sensitive to ice melting assumptions, with lower SLR reducing losses but not eliminating the need for adaptation. When comparing the benefits of planned adaptation under different reference points (no-adaptation vs. autonomous adaptation), the benefits are significantly smaller when autonomous adaptation is included. This shows the bias introduced by the no-adaptation assumption. The comparison of planned adaptation versus autonomous adaptation alone shows that in certain regions, particularly in Europe, autonomous adaptation is comparatively more effective, which can be attributed to the higher investment requirements of planned protection measures. The cost-effectiveness of the combined adaptation strategy is notably pronounced in regions such as Southeast Asia and Oil-Exporting Regions, where protection measures tend to be less widespread due to lower population density, reducing capital requirements.

These findings align with other studies using similar soft-linked CGE approaches, but direct comparisons are limited by variations in cost assumptions and modeling approaches. The aggregate focus on GDP, however, masks distributional effects between public and private sectors. Publicly financed protection can reduce the provision of other public goods, while private adaptation capacity might be insufficient for funding autonomous actions. The model's simplification of autonomous adaptation to migration alone is a limitation, but it reveals a bias in assessing protection benefits. The chosen 1-in-1-year flood return level threshold for migration may represent a limit case and further research is needed to better understand migration dynamics and their cost. Additional limitations include excluding costs from erosion, salinization, ecosystem changes, and intangible losses. Also, potential co-benefits from adaptation are omitted. Although the study analyzes the time horizon to 2050, SLR is projected to increase beyond this period, so the macroeconomic losses are expected to increase further.

This study demonstrates the importance of incorporating both protection and migration in macroeconomic assessments of SLR. Combining these strategies can reduce costs compared to protection alone. Autonomous adaptation, particularly migration, significantly lowers losses, especially in developing regions. The overestimation of protection benefits in existing literature arises from the use of unrealistic no-adaptation reference scenarios. Future studies should consider autonomous adaptation in reference scenarios for accurate cost-benefit analysis. The dominant role of accumulated macroeconomic effects over time necessitates immediate adaptation. Future research should refine the modeling of autonomous adaptation, incorporate more direct costs and co-benefits, and investigate distributional effects.

The study's limitations include the simplification of autonomous adaptation to only migration, which might not fully capture the diversity of adaptation strategies. The model excludes several direct costs of SLR, such as erosion, salinization, and ecosystem changes, as well as intangible costs like loss of culture. Furthermore, the assumption of internal migration only, and the assumed neutrality of changes in spatial economic agglomeration, might introduce some degree of bias into the results. Finally, the study focuses on macroeconomic effects at an aggregate level, neglecting distributional impacts between various actors and socioeconomic groups. The model does not explicitly account for the co-benefits of planned adaptation, such as the creation of public spaces.