In total, the 24 identified sources cover investment needs estimates for 9 SDGs. Three of them cover targets related to multiple SDGs, while the rest focus on specific SDGs or SDG-relevant sectors such as healthcare, education, infrastructure, or energy. To analyse and compare the estimates, it is important to consider which SDG target(s) are considered, which costs are estimated, and how. Seven sources present investment needs related to SDG7 (access to energy). Five sources were found for SGD 6 (clean water and sanitation), four each for SDG 13 (climate action), SDG 3 (good health and well-being) and SDG 9 (infrastructure), three for SDG 4 (quality education), two for SDG 2 (zero hunger) (Table 1), and one for SDGs 14 and 15 together (protecting biodiversity under water and on land).

The methods used in the identified literature, where specified, are presented using the categorisation in ST15 into simple unit cost estimates, intervention-based needs assessments, computational general equilibrium (CGE) models, incremental capital-output ratio approaches (but no studies were found that applied this method), or Integrated Assessment Models (IAMs).

Unit cost method estimates are typically based on available investment data for a specific intervention (e.g. the costs of building a school or sanitation facility). It can also be derived from current government spending divided by the actual progress on key indicators. As such, past expenditure is used to estimate the investment needed to meet the targeted indicator level. Unit cost methods are relatively simple estimation tools useful to make first-order estimates. However, there are questions about whether one can scale investment needs under different pathways using this method. Intervention-based needs assessments apply a similar method but account for dynamically evolving capital and operational costs of implementing a specified set of interventions. They benefit from comprehensive coverage of assessed needs with better identification of overlaps and gaps disaggregated by population or region and an assessment of differentiated responses addressing the differentiated capabilities and needs. CGE models represent an economy in equilibrium using stylised production functions for SDG-relevant economic sectors such as healthcare, education, or energy production. Dynamic economy-wide effects of targeted investments, such as through changing prices, can be assessed by varying the input investment levels and combinations in the modelled sectors. Finally, studies based on IAMs can leverage the detailed representation of various relevant but complex environmental, economic and societal systems and provide investment estimates accounting for cross-sectoral effects under different policies and scenarios.

Other methodological details and underlying assumptions are listed in the supplementary information (Table B in S1 Table).

Fig 1 presents the overview of published estimates. All reported estimates are presented in USD2019 using the World Bank Consumer Price Index [43], unless otherwise specified. The numbers refer to the investment needs additional to baseline, i.e. incremental needs, except SDG 9 (covering transport and communication infrastructure), for which only total investment needs were found in the literature. The latter results are reported separately in Fig 1 given their significance and the lack of viable alternatives. All estimates are disaggregated to specific targets whenever possible. In case such disaggregation is not possible based on the information presented in the sources, the estimates are mapped directly against the associated goal. However, very few studies cover the needs of all targets defined for a goal. Estimates reported as a single value are plotted as dots, while a line plots those reported as ranges. Estimates from the previous ST15, and WIR14 studies (as reported in the SDG Investment Trends Monitor 2019 [8]) are plotted in grey for comparison.

Recent strides reflect significant progress towards achieving SDG 3 (good health and well-being) targets overall, particularly towards reduced maternal and neonatal mortalities. Still, less than half of the global population is estimated to have access to essential health services [44]. The grim outlook is further exacerbated by the strain on existing healthcare capacity, program disruptions (such as immunisation drives) and resource re-allocation from the ongoing pandemic.

Investment estimates in the healthcare sector follow a unit costs approach that can address regionally relevant needs to bridge the gap to global benchmarks [45]. A 2017 WHO study estimates an annual investment gap of USD 400 billion for comprehensive health service delivery in 67 low- and low-to-middle-income countries [21]. It’s simple, bottom-up unit costs approach covers the costs of scaling up WHO recommended interventions. They primarily address access to universal healthcare services (SDG 3.8) but also contribute to other SDG 3 targets such as maternal mortality (SDG 3.1), neonatal health (SDG 3.2), epidemics and communicable diseases, including AIDS (SDG 3.3), access to reproductive healthcare (SDG 3.7).

Stenberg et al. [22] cover a smaller set of direct SDG 3 targets and report an annual investment gap of USD 257 billion. This estimate includes associated investment needs beyond the health sector. For instance, investments directly related to public health interventions include investments into diagnostic tests, orthopaedic devices, hygiene and sanitation infrastructure. These additional inclusions are associated with its broader focus on primary healthcare contributing to universal health coverage.

Bertram et al. [23] focus on SDG 3.4 (non-communicable diseases) related interventions, specifically those combating cardiac diseases in the top twenty countries with the highest risk prevalence. At USD 120 billion per year, the study estimates a relatively high investment need for the subset of the targets covered by the previous studies, likely due to higher costs in developed economies where cardiac diseases are more prevalent.

Investments in air pollution control addressing SDG 3.9 are estimated by McCollum et al. [29] under the global emissions scenarios. For scenarios consistent with the Paris Agreement and SDG 13, they estimate annual investments up to USD 669 billion.

All the above estimates are substantially higher than the estimates of around 100 USD annually, as reported by ST15 and WIR14.

SDG 4 deals with inclusive and equitable access to quality education. It includes early childhood care and pre-primary targets to primary and secondary education and tertiary and vocational education for decent employment.

The Global Education Monitoring Report [24] performs a comprehensive costing exercise for the investments needed to meet all SDG 4 targets, including school infrastructure (classrooms, materials, etc.), hiring and training of teachers, development of quality curricula, policy costs such as those for mid-day meals, incentivising enrolment, WASH, promoting inclusivity across gender, disability, ethnicity, and socio-economic status. It calculates the annual investments needed at USD 340 billion for low- and low-to-middle-income countries, leading to a financing gap of USD 39 billion over baseline domestic funding. Another UNESCO report clarifies that updated data and rechecking model assumptions increase the total investments needed to USD 504 billion, and the gap from USD 39 billion to USD 148 billion. It further estimates that the COVID-19 pandemic likely increases this gap by up to USD 45 billion to USD 193 billion [25]. The increase is driven by reduced domestic budgets triggered by economic recession and the additional interventions needed to overcome the effects of the global pandemic.

The study by McCollum et al. [29] uses UNESCO’s data on average costs per student. It estimates that USD 194 billion per year will be needed to achieve projected primary and secondary education demands, reflecting only infrastructure investments needed.

In comparison, ST15 estimates an additional annual investment need towards education targets at USD 213 billion, while the WIR14 estimates at a comparable USD 274 billion.

Goal 6 of the SDGs aims for equitable access to water and sanitation (WASH), which is acknowledged as a basic human right being a pre-requisite for survival [46]. The targets under this goal focus on sustainable management of water resources in addition to the actual provision of drinking water, hygiene and sanitation services. These include ensuring water quality and limiting pollution, transboundary water management, improving water use efficiency and limiting unsustainable withdrawals.

Five additional sources estimating the investment needs for this SDG were found. The estimates vary widely across the studies, from about USD 30 billion to USD 1.1 trillion per year.

Hutton & Varughese [26] estimate that providing universal access to basic WASH services poses an investment need between USD 15 and 50 billion (mean USD 32 billion) annually additional to baseline spending, and up to USD 179 billion per year additional to the baseline to reach a higher “safely-managed WASH” ambition level. The estimates are derived using a simple unit costs approach to include community wells and latrines, access to hand-washing stations with soap for basic WASH, and piped water and faecal waste treatment for safely-managed WASH.

The OECD estimates a much higher annual investment need of almost USD 0.9 trillion (of which USD 482 billion within their member states). This estimate includes the investments needed to maintain and upgrade urban water and sanitation infrastructure [27]. In addition to providing access to drinking water and sanitation services, much higher investments are needed towards broader water resource management infrastructure to ensure sustainable extraction of this limited resource, meeting projected global and local water demands. It is, however, unclear if the presented estimates total investment needs or additional to a baseline.

The above estimates do not take the impact of climate change into account. The IAM-based studies that have estimated investment needs for SDG 6 take climate change into account, primarily through the projected water demand in alternative climate, energy-use, behavioural and population growth scenarios. Apart from the water withdrawal and distribution infrastructure associated with meeting increased demand, they also account for the consequent infrastructure needs to ensure water supply (e.g., pollution control, wastewater recycling and desalination), to reduce water demand (e.g. towards improving household, agricultural, and industrial water-use efficiency), and for integrated water resource management. Total annual investment needs estimates to meet SDG 6 targets derived by this method range from USD 445 to 885 billion [28,29].

A World Resources Institute (WRI) working paper provides a comprehensive needs estimate for SDG 6 targets. It includes full capital and operational expenditures spanning access to drinking water and sanitation services, agricultural and industrial water pollution control, reducing scarcity, and regulation and integrated water management [30]. At 1.12 trillion USD, it is the highest estimated investment gaps to reach SDG 6 and one of the highest for any SDG.

ST15 focuses only on basic WASH access targets in SDG 6, and its estimated investment gap of USD 46–50 billion is comparable to the recent estimates that looked at similar targets. The UNCTAD estimate of USD 285 billion in WIR14 still focuses mainly on WASH targets but accounts for more ambitious levels of WASH and pollution control infrastructure.

SDG 7 aims to achieve universal, reliable, and affordable access to modern energy services. As of 2018, 789 million people worldwide still lacked access to electricity, and 2.8 billion–one-third of the planet’s population–to clean cooking stoves and fuels. Increasing the share of renewables remains a challenge both in existing and planned energy infrastructure deployment. The greatest deficit is seen in Sub-Saharan Africa (SSA), with about 548 million people, or over half the total population in the region still lacking access to electricity.

Investment needs estimates to bridge these enormous gaps are consequently predominantly focused on the SSA region. Interventions in electricity access are modelled by increasing tiers of the services provided by degrees of access ranging from enough electricity to light a few light bulbs and a cellphone to full 24x7 access to grid electricity. Investment estimates further differ from each other by the ambition levels and technological choices, and energy sources considered in each tier.

In the State of Electricity Access Report 2017, the World Bank estimates an annual investment need of USD 50 billion for full 24x7 grid electricity access globally [35]. The IEA estimates an annual investment need of USD 102 billion for the SSA region with a mix of full-grid, mini-grid, and stand-alone transmission systems [34]. However, the methods and data sources used are not made explicit in either of the above studies.

Dagnachew et al. [31] and Lucas et al. [32] used the IMAGE integrated assessment model to estimate the annual capital investments needed to achieve universal access to electricity in SSA by 2030 at about USD 10 to 40 billion and USD 51 billion, respectively. Both estimated these investment needs in addition to a USD 21 billion annual baseline investment. These recent estimates are notably lower than those estimated by ST15 and WIR14, since the transformational action needed in the energy supply systems are now largely accounted for under SDG 13 (climate action).

Hof et al. [33] use the same IAM framework to estimate that USD 54 billion would be needed per year towards modern stovetops and associated fuels in the same SSA region in the same period to reach universal access to clean cooking solutions. However, the multi-IAM approach in McCollum et al. [29] calculates an estimated USD 135 billion to USD 350 billion in capital and fuel costs to achieve universal access to clean cooking globally.

For comparison, ST15 estimates an additional investment of USD 72 billion to USD 95 billion globally to achieve this target, attributable largely to the little progress made over the last two decades outpaced by high population growth in the vulnerable regions [44].

SDG 9 has three objectives: building resilient infrastructure, developing sustainable industrialisation, and fostering innovation. Investment needs estimates in the reviewed literature related to this goal are invariably associated with infrastructure needs alone since infrastructure development is well suited for measurable goal-based modelling of interventions and typically involve enormous investments. Moreover, infrastructure is recognised to influence multiple targets spanning across all 17 SDGs directly or indirectly [47].

All investment needs that were found in the literature refer to total needs instead of additional to baseline. Since they also vary in geographic coverage, costing methodology, and assumptions, it is impossible to apply a common baseline exogenously here.

Investment needs assessments in this sector are available for four categories: transport infrastructure, telecommunication infrastructure, energy infrastructure, and infrastructure related to providing basic services such as healthcare, water, food, or education. To avoid double-counting, the findings of the first two categories of transport and telecommunication infrastructure are presented under SDG 9. The estimates related to the other categories are accounted for in the respective SDGs associated with those sectors.

A recent assessment is made by Fisch-Romito & Guivarch [39], who use the Imaclim-R IAM to estimate global transportation infrastructure investment needs in low and high carbon scenarios at USD 1–4 trillion per year. The study also presents additional sources in literature for comparison. Annual investment needs estimates by Dulac [37] at USD 2 trillion, and from the Energy technology Perspective, 2016 [38] are pegged at USD 2.7 trillion. An OECD study updates the latter estimate by IEA to include investments into airports and ports, bringing the total estimate to USD 2.9 trillion annually [48].

Rozenberg & Fay [36] focus on low- and low-to-middle-income countries and use an IAM based method to estimate an annual investment need of USD 63 billion in a socio-economic scenario marked by increasing urban density, high utilisation rates of public transport, and efficient returns to investments.

Total investment needs as reported by WIR14 amounted to USD 350 to 770 billion, of which USD 300 billion as baseline investment and USD 50 billion to USD 470 billion as additional need. The recent studies, therefore, seem to report much higher numbers than WIR14.

McCollum et al. [29] and Bertram et al. [40] offer the most comprehensive overview so far of the investments in the energy systems required to mitigate climate change to 1.5°C (SDG 13).

The McCollum et al. study contrasts the Paris-consistent 1.5°C scenario with reference scenarios representing continuation of current policies, and trends defined by the nationally determined contributions (NDCs) to climate mitigation efforts. The resulting annual investment gap between the 1.5°C and current policy scenarios is estimated at USD 900 billion, ranging from USD 481 billion to USD 1.6 trillion, and covering expenditures towards increasing the share of renewables substantially and doubling the energy efficiency improvement rate. The study uses a multi-model comparison from six IAMs that model the global energy/economy system using diverse sectoral coverage and methodologies. This breadth allows calculating how such an energy system transformation by 2030 would affect the investment needs of other SDG targets from air pollution, clean cooking fuels, water and sanitation, food security, and education. These investment needs are presented alongside other estimates for the respective SDGs.

Bertram et al. [40] also use a similar multi-model approach using IAMs, but feature a substantially lower global emission budget and higher targeted annual reductions compared to the previous study. The approach highlights the critical early- to mid-term emission reductions, mainly through investments in improving energy efficiency in highly energy intense industries and decarbonising energy sources, necessary to exceed peak 1.5°C warming. Capital heavy energy transition investments are thus front-loaded in the near term compared to approaches which rely on late term atmospheric carbon dioxide removal to meet end of century warming targets. The study consequently pegs the estimated annual investment gap at USD 300 billion–USD 3 trillion.

Very little information can be found on the investment needs towards climate adaptation at the global scale. One study, however, addresses a known form of adaptation to one of the better understood global climate change impacts–sea-level rise. It estimates the annual investments needed to build dykes to prevent coastal flooding globally in response to the projected sea-level rise at about USD 160 billion [41].

While SDGs 14 and 15 aim to protect natural environments, respectively under water and on land, only SDG 14 presents a numerical target: to protect 10% of the marine areas. Growing evidence suggests that at least 30% of land and water surface should be protected to avoid the collapse of natural environments [49,50]. Accordingly, the 15th Conference of Parties of the Convention on Biological Diversity defined 30% as the target for the extent of protected areas by 2030 [51]. Such target is also more amenable to cumulative investment needs assessment compared to adding up the investment needs for conservation of individual endangered species [52].

Waldron et al. [42] estimated investments needed to increase protected areas from 2019 levels up to 30% each of land and water surfaces by 2030. Cost estimations are based on data on conservation spending covering 97 countries, mostly drawn on the financial scorecards submitted to the UNDP on protected areas system needs [53]. The study developed a baseline scenario with no expansion of protected areas. Six additional scenarios were also developed to implement different potential locations of the expanded protected areas based on degree of biodiversity protection, as well as political and economic considerations by the authors. The scenarios provided a corresponding range for additional investment needs from USD 85 to 166 billion annually to reach the 30% protected areas target.

Although beyond the scope of the present study, it is important to note here that the opportunity costs involved with implementing conservation activities are often higher than the direct implementation costs. An analysis presented in a recent Convention on Biological Diversity report suggested that at USD 638 billion to USD 876 billion annually, estimates including the opportunity costs incurred are an order of magnitude higher than the direct investment needs [54].

WIR14 does not present investment needs towards biodiversity SDGs. ST15 presents a source for additional annual investment needs at USD 18 to 48 billion, but only for the period 2014–2018 and thus not relevant for comparison in the present study.