Non-linear interlinkages and key objectives amongst the Paris Agreement and the Sustainable Development Goals
PPublished as a workshop paper at ICLR 2020 N ON - LINEAR INTERLINKAGES AND KEY OBJECTIVESAMONGST THE P ARIS A GREEMENT AND THE S USTAIN - ABLE D EVELOPMENT G OALS
Felix Laumann
Department of MathematicsImperial College London [email protected]
Julius von K ¨ugelgen
Max Planck Institute for Intelligent Systems, T¨ubingenDepartment of Engineering, University of Cambridge
Mauricio Barahona
Department of MathematicsImperial College London A BSTRACT
The United Nations’ ambitions to combat climate change and prosper human de-velopment are manifested in the Paris Agreement and the Sustainable Develop-ment Goals (SDGs), respectively. These are inherently inter-linked as progress to-wards some of these objectives may accelerate or hinder progress towards others.We investigate how these two agendas influence each other by defining networksof 18 nodes, consisting of the 17 SDGs and climate change, for various groupingsof countries. We compute a non-linear measure of conditional dependence, thepartial distance correlation, given any subset of the remaining 16 variables. Thesecorrelations are treated as weights on edges, and weighted eigenvector centralitiesare calculated to determine the most important nodes.We find that SDG 6, clean water and sanitation , and SDG 4, quality education ,are most central across nearly all groupings of countries. In developing regions,SDG 17, partnerships for the goals , is strongly connected to the progress of otherobjectives in the two agendas whilst, somewhat surprisingly, SDG 8, decent workand economic growth , is not as important in terms of eigenvector centrality.
NTER - LINKED HUMAN AND NATURAL WORLDS
The state-of-the-art in sustainability is described by two United Nations (UN) landmark agendas, theParis Agreement (UN, 2015a) and the Sustainable Development Goals (SDGs) (UN, 2015b). Whilstthe former focuses on preventing a global climate crisis with far reaching consequences by limitingglobal warming to 1.5 to 2 ◦ C above pre-industrial levels, the purpose of the latter is to end poverty,protect the planet and ensure that all people enjoy peace and prosperity by 2030. Any action for theprogress on either agenda often has an influence on the other (UN Climate Change, 2019), reflectingthe complexity of the human and natural worlds.This inter-linked nature gives rise to opportunities for the creation of synergistic interventions: civil,corporate and institutional actions can efficiently create impact across both agendas, thereby im-proving the world profoundly. On the other hand, this inter-linked construct can also be subjectto trade-offs between objectives, i.e., progress towards one agenda constrains progress towards theother. In this work, we aim to discover how climate change, as measured by local temperature rises,and the 17 SDGs are inter-linked by learning the structure of undirected graphs over these variablesfrom their (conditional) dependencies.Adding climate change as an 18 th variable is motivated by the observation that temperature rises (orany other direct metrics of climate change) are not actually tracked within SDG 13 ( climate action ).Indicators of SDG 13 only track inputs (such as investment), means (such as plans and strategies),1 a r X i v : . [ ec on . E M ] A p r ublished as a workshop paper at ICLR 2020and impacts (number of people affected by disasters), but they do not account for outputs, such aschanges in temperature or green house gas emissions. We use distance correlation (Sz´ekely et al., 2007) as a measure of non-linear dependence betweenvariables of possibly varying dimensions. To account for possible interactions, each pair of variablesis conditioned on any subset of the remaining variables, and the minimum resulting distance corre-lation is taken as the weight on an edge between these two variables. Subsequently, the weightedeigenvector centrality of every node is calculated to measure its importance within the network.In summary, the contributions of this paper include: first, the application of a non-linear measureof (conditional) dependence to SDG data, thereby relaxing the linearity assumption on the nature ofinterlinkages between the SDGs, compared to the work of Lusseau & Mancini (2019); and secondly,the use of eigenvector centrality as a relative measure which also takes the importance of a node’sneighbours into account, as opposed to simple degree centrality as used by McGowan et al. (2019).
ETHODOLOGY
We use data provided by the World Bank (2020b) and the UN (2020) in form of time-series forvarious indicators, which measure progress towards their associated SDGs, in conjunction withtemperature recordings (World Bank, 2020a). In total, these three sources provide 379 time-series, which are available on a country-level with annual measurements from 2000 to 2016 .Apart from measurements for the 17 SDGs, we introduce climate change as an additional vari-able which we define by annual average temperature per country. We consider these 18 variablesas the set of nodes V of an undirected graph G . We learn the graph structure by computing par-tial distance covariances (Sz´ekely et al., 2014) between any pair ( X, Y ) of nodes, given any subset Z ⊆ V \ ( X, Y ) of the remaining 16 nodes. This yields a sparsely-connected undirected graphwith weighted edges E capturing non-linear dependencies between variables. Using these weights,we compute weighted eigenvector centralities (Newman, 2018, p.159; Appendix A.2) to find themost important nodes. Code to reproduce our findings and visualisations of networks may be foundonline at https://github.com/felix-laumann/SDG-dataset .2.1 D ISTANCE COVARIANCE
Let X ∈ R d X and Y ∈ R d Y be two random vectors with finite first moments, i.e., E [ X ] , E [ Y ] < ∞ .The distance covariance between X and Y , denoted by V ( X, Y ) , is a measure of dependencebetween X and Y with the following important properties: (i) V ( X, Y ) ≥ , with equality if andonly if X and Y are independent, i.e., it is a non-parametric measure that—unlike, e.g., standardcorrelation—is able to pick up complex non-linear dependencies; (ii) V ( X, Y ) = V ( Y, X ) , i.e.,it is symmetric; and (iii) unlike many other dependence measures V ( X, Y ) is well-defined evenfor d X (cid:54) = d Y . This last point makes it particularly useful for our setting where, due to the differentnumbers of indicators per SDG, dimensionality varies considerably between variables.Formally, the distance covariance between X and Y is defined as V ( X, Y ) = (cid:107) f X,Y ( t, s ) − f X ( t ) f Y ( s ) (cid:107) := (cid:90) | f X,Y ( t, s ) − f X ( t ) f Y ( s ) | w ( t, s ) dt ds (1)where w ( t, s ) := ( | t | d X d X | s | d Y d Y ) − , and where the characteristic function f of a random variable Z is denoted as f Z ( t ) = E [ e itZ ] with i = − .The corresponding distance correlation R is the normalised distance covariance, computed by R ( X, Y ) = (cid:40) V ( X,Y ) √ V ( X,X ) V ( Y,Y ) , if V ( X, X ) V ( Y, Y ) > , if V ( X, X ) V ( Y, Y ) = 0 . (2)Properties of R include: (i) ≤ R ( X, Y ) ≤ ; and (ii) R ( X, Y ) = 1 if and only if there existsa vector a , a non-zero real number b , and an orthogonal matrix C such that Y = a + b C X . Only recently (and after performing the present analysis) have ”total greenhouse gas emissions” beenadded as an output-quantifying indicator (13.2.2). For detailed descriptions of indicators, see https://sustainabledevelopment.un.org/sdgs We impute missing values (especially for the time 2000-2005) using a weighted average across countries(where data is available) with weights inversely proportional to the Euclidean distance between indicators. left ), and between Western Asia and Northern Europe ( right ). Results for all groupings can befound in B.2.
Global South Global NorthSDG 6 0.48 SDG 6 0.43SDG 4 0.42 SDG 4 0.40SDG 7 0.38 SDG 9 0.33SDG 17 0.37 SDG 3 0.32SDG 3 0.26 SDG 17 0.29SDG 15 0.25 SDG 7 0.27 Western Asia Northern EuropeSDG 6 0.48 SDG 4 0.38SDG 4 0.36 SDG 3 0.35SDG 17 0.34 SDG 6 0.30SDG 3 0.33 SDG 16 0.30SDG 16 0.32 SDG 7 0.29SDG 7 0.26 SDG 9 0.28
Since V ( X, Y ) and therefore R ( X, Y ) are defined in terms of the underlying joint distribution of ( X, Y ) which is usually not known, we require a way to estimate them from data. Definitions ofbiased and unbiased estimators, referred to as A and ˜ A , can be found in Appendix A.1.1 and A.1.2.2.2 P ARTIAL DISTANCE COVARIANCE
As we deal with graphs of 18 nodes, any pairwise covariance may occur through the remaining16 nodes. Thus, we condition any pair of nodes ( X, Y ) on any subset Z ⊆ V \ ( X, Y ) of theremaining 16 nodes. The pairwise distances c ij = (cid:107) Z i − Z j (cid:107) and the distance matrix C ij for Z are computed equivalently to A ij and B ij for X and Y as explained in Appendix A.1.1. For anynumber n ≥ of samples { ( x i , y i , z i ) } ni =1 from ( X, Y, Z ) , we define a Hilbert space H n overdistance matrices computed on these n points, with inner products (cid:104)· , ·(cid:105) as defined in AppendixA.1.2 (Sz´ekely et al., 2014). With this, we can compute partial distance covariances for randomvectors of varying dimensions as follows.Let ˜ A ( x ) , ˜ B ( y ) and ˜ C ( z ) be elements of the Hilbert space H n corresponding to the distance ma-trices computed using the samples x = ( x , ..., x n ) , y = ( y , ..., y n ) , and z = ( z , ..., z n ) , re-spectively. The projection P z ( x ) of ˜ A ( x ) onto ˜ C ( z ) and the complementary orthogonal projection P z ⊥ ( x ) are defined by P z ( x ) := (cid:104) ˜ A ( x ) , ˜ C ( z ) (cid:105)(cid:104) ˜ C ( z ) , ˜ C ( z ) (cid:105) ˜ C ( z ) , and P z ⊥ ( x ) := ˜ A ( x ) − P z ( x ) = ˜ A ( x ) − (cid:104) ˜ A ( x ) , ˜ C ( z ) (cid:105)(cid:104) ˜ C ( z ) , ˜ C ( z ) (cid:105) ˜ C ( z ) , (3)respectively. The sample partial distance covariance is then defined as V n ( X, Y | Z ) = (cid:104) P z ⊥ ( x ) , P z ⊥ ( y ) (cid:105) = n ( n − (cid:80) ni (cid:54) = j (cid:0) P z ⊥ ( x ) (cid:1) ij (cid:0) P z ⊥ ( y ) (cid:1) ij . (4)Finally, we can normalise these covariances to arrive at the sample partial distance correlations R n ( X, Y | Z ) = (cid:40) (cid:104) P z ⊥ ( x ) ,P z ⊥ ( y ) (cid:105)(cid:107) P z ⊥ ( x ) (cid:107) (cid:107) P z ⊥ ( y ) (cid:107) , if (cid:107) P z ⊥ ( x ) (cid:107) (cid:107) P z ⊥ ( y ) (cid:107) (cid:54) = 00 , if (cid:107) P z ⊥ ( x ) (cid:107) (cid:107) P z ⊥ ( y ) (cid:107) = 0 , (5)which serve as weights on edges between any two nodes. ESULTS
We apply this methodology to the data set of the aforementioned 379 indicators for various groupingsof countries, for which countries are assumed to be independent samples. This assumption allowsus to see the indicators’ non-stationary time-series as d -dimensional probability distributions, where d = indicators × years. Whilst we only describe the networks of a few groupings in this section,we would like to refer to Appendix B for results on all groupings.Firstly, we compare the Global South and the Global North (see Figure 1). The accompanied eigen-vector centralities are shown in Table 1. In both groupings, SDG 6, clean water and sanitation ,followed closely by SDG 4, quality education , are the most central objectives of the 18 variables.In the Global South, temperature rises are more strongly dependent on variables than in the GlobalNorth, which broadly aligns with King & Harrington (2018) who find that geographical areas in3ublished as a workshop paper at ICLR 2020Figure 1: Networks with weighted edges of ( left ) the Global South and ( right ) the Global North. Theminimum partial distance correlations between the two adjacent nodes X and Y , given any subset Z ⊆ V \ ( X, Y ) are weights on edges.the Global South are more vulnerable to climate change than regions in the Global North. Further,SDG 1, no poverty , is strongly linked to SDG 14, life below water , in the Global South. This maybe explained by the dependence of small island developing states (SIDS)—all of which lie in theGlobal South—on marine life to provide for their citizens’ living.Contrarily, the Global North strongly depends on SDG 9, industry, innovation and infrastructure , tomaintain its citizens’ high levels of living standards and to further progress towards other SDGs, aswell as climate change mitigation and adaptation. Moreover, SDG 7, clean and affordable energy ,is closely related to SDG 15, life on land , which could result from the increasing area of biodiverseland populated by wind turbines, solar panels, or water dams (e.g., Hernandez et al., 2015).Next, we compare two geographical regions, Western Asia and Northern Europe, shown in Figure 2with accompanied eigenvector centralities in Table 1. In Western Asia, SDG 6 together with SDG4 are again the two most central nodes, but SDG 16, peace, justice and strong institutions , is alsoimportant, likely to be associated with the unstable political circumstances in this area during theperiod of recorded measurements. Additionally, SDG 5, gender equality , is strongly linked to SDG17, partnerships for the goals , which coincides with the remarkably low percentage of women inmanagerial positions in Western Asia. In contrast, Northern Europe does not see a remarkable difference between the centralities of SDGs6 and 4 to all others, but finds SDGs 4, 6, 3, and 17 with almost equivalently high centralities. AsFigure 2: Networks with weighted edges of ( left ) Western Asia and ( right ) Northern Europe. Theminimum partial distance correlations between the two adjacent nodes X and Y , given any subset Z ⊆ V \ ( X, Y ) are weights on edges. In Saudi Arabia, for example, only 5 to 9% of managerial positions were held by women from 2000 to2015, whereas this number fluctuated between 32 and 36% in the United Kingdom in the same period (UN(2020), indicator 5.5.2) industry, innovation and infrastructure are of particular importance to progresstowards the SDGs, and we fine that clean and affordable energy is closely linked to life on land .We note, however, that most edges found in our network analysis are not statistically significant at p = 0 . , using the test of Sz´ekely et al. (2014). This is likely linked to the high dimensionality ofthe data and the short recording period. The present work is thus only a first step, and further analysisis needed to better understand non-linear interlinkages between the SDGs and climate change. ONCLUSIONS
We report findings of our work in progress towards discovering dependencies amongst the Sustain-able Development Goals (SDGs) and climate change. As a first step, we compute partial distancecorrelations between the 17 SDGs and climate change, as measured by indicators associated to theSDGs and annual average temperature, respectively. Using these measurements of non-linear de-pendence as edge weights in a network over these variables, we determine eigenvector centralitiesto unveil which variables are of particular importance, given the available data. Our results indicatethat SDG 6, clean water and sanitation , together with SDG 4, quality education , are the most centralnodes in nearly all continents and other groupings of countries. In contrast to many contemporarypolicies, our preliminary results suggest that economic growth , as measured by SDG 8, appears notto play as central of a role for sustainable development or mitigating climate change as other SDGs. R EFERENCES
Rebecca R Hernandez, Madison K Hoffacker, Michelle L Murphy-Mariscal, Grace C Wu, andMichael F Allen. Solar energy development impacts on land cover change and protected areas.
Proceedings of the National Academy of Sciences , 112(44):13579–13584, 2015.Andrew D King and Luke J Harrington. The inequality of climate change from 1.5 to 2 c of globalwarming.
Geophysical Research Letters , 45(10):5030–5033, 2018.David Lusseau and Francesca Mancini. Income-based variation in sustainable development goalinteraction networks.
Nature Sustainability , 2(3):242–247, 2019.Philip JK McGowan, Gavin B Stewart, Graham Long, and Matthew J Grainger. An imperfect visionof indivisibility in the sustainable development goals.
Nature Sustainability , 2(1):43–45, 2019.Mark Newman.
Networks . Oxford University Press, 2018.G´abor J Sz´ekely, Maria L Rizzo, Nail K Bakirov, et al. Measuring and testing dependence bycorrelation of distances.
The Annals of Statistics , 35(6):2769–2794, 2007.G´abor J Sz´ekely, Maria L Rizzo, et al. Partial distance correlation with methods for dissimilarities.
The Annals of Statistics , 42(6):2382–2412, 2014.UN. Paris Agreement. http://unfccc.int/files/essential_background/convention/application/pdf/english_paris_agreement.pdf , 2015a.UN. The 2030 Agenda for Sustainable Development. https://sustainabledevelopment.un.org/content/documents/21252030%20Agenda%20for%20Sustainable%20Development%20web.pdf , 2015b.UN. Sustainable Development Goals indicators. https://unstats.un.org/sdgs/indicators/database/ , 2020. Accessed: 2020-01-28.UN Climate Change. Global Conference on Strengtheining Synergies, ConferenceSummary. https://sustainabledevelopment.un.org/content/documents/25256WEB_version.pdf , 2019.World Bank. Climate Change Knowledge Portal. https://climateknowledgeportal.worldbank.org/download-data , 2020a. Accessed: 2020-01-28.World Bank. Sustainable Development Goals. https://datacatalog.worldbank.org/dataset/sustainable-development-goals , 2020b. Accessed: 2020-01-28.5ublished as a workshop paper at ICLR 2020
A A
PPENDIX A A.1 D
ISTANCE COVARIANCE ESTIMATORS
A.1.1 B
IASED ESTIMATORS
Suppose that we have access to a sample of pairs ( x , y ) , ..., ( x n , y n ) i.i.d ∼ P X,Y . First, define the pairwise distances : a ij := (cid:107) x i − x j (cid:107) and b ij = (cid:107) y i − y j (cid:107) ∀ i, j = 1 , ..., n . Next, define thecorresponding distance matrices , denoted by ( A ij ) ni,j =1 and ( B ij ) ni,j =1 , as follows: A ij = (cid:40) a ij − n (cid:80) nl =1 a il − n (cid:80) nk =1 a kj + n (cid:80) nk,l =1 a kl , if i (cid:54) = j , if i = j (6)and B ij = (cid:40) b ij − n (cid:80) nl =1 b il − n (cid:80) nk =1 b kj + n (cid:80) nk,l =1 b kl , if i (cid:54) = j , if i = j. (7)Having computed these, the sample distance covariance V n ( X, Y ) can be estimated by V n ( X, Y ) = 1 n n (cid:88) i,j =1 A ij B ij , (8)which converges almost surely to the population distance covariance V ( X, Y ) as n → ∞ (Sz´ekelyet al., 2014).A.1.2 U NBIASED ESTIMATORS
Unbiased estimators of the distance covariance are denoted as Ω n ( x, y ) . Firstly, we must redefineour distance matrices ( A ij ) ni,j =1 and ( B ij ) ni,j =1 , which we call ( ˜ A ij ) ni,j =1 and ( ˜ B ij ) ni,j =1 as ˜ A ij = (cid:40) a ij − n − (cid:80) nl =1 a il − n − (cid:80) nk =1 a kj + n − n − (cid:80) nk,l =1 a kl , if i (cid:54) = j ;0 , if i = j (9)and ˜ B ij = (cid:40) b ij − n − (cid:80) nl =1 b il − n − (cid:80) nk =1 b kj + n − n − (cid:80) nk,l =1 b kl , if i (cid:54) = j ;0 , if i = j. (10)Finally, we can compute the unbiased estimator Ω n ( X, Y ) for V ( X, Y ) as the dot product (cid:104) ˜ A, ˜ B (cid:105) : Ω n ( X, Y ) = (cid:104) ˜ A, ˜ B (cid:105) = 1 n ( n − n (cid:88) i,j =1 ˜ A ij ˜ B ij (11)A.2 E IGENVECTOR CENTRALITY
For any graph G := ( V, E ) , let K be the adjacency matrix of graph G with k v,t equal to the weighton the edge between node v and t . The eigenvector centrality x of node v is a measure relative to allother nodes in G , defined as x v = 1 λ (cid:88) t ∈G k v,t x t , (12)where λ is the greatest eigenvalue in the eigenvector equation K x = λ x , subject to x (cid:54) = 0 . Conse-quently, this centrality measure is an extension of the widely used degree centrality by consideringthe centrality of its neighbours besides its own. 6ublished as a workshop paper at ICLR 2020 B A
PPENDIX B B.1 N
ETWORKS OF GROUPINGS
IGENVECTOR CENTRALITIES B . G R OU P I NG S O F C OUN T R I E S N o r t h e r n A fr i ca E a s t e r n A fr i ca M i dd l e A fr i ca S ou t h e r n A fr i ca W e s t e r n A fr i ca S ub - S a h a r a n A fr i ca A fr i ca C a r i bb ea n C e n t r a l A m e r i ca S ou t h A m e r i ca L a ti n A m e r i caa nd t h e C a r i bb ea n N o r t h A m e r i ca A m e r i ca s C e n t r a l A s i a E a s t e r n A s i a S ou t h - ea s t e r n A s i a S ou t h e r n A s i a W e s t e r n A s i a A s i a E a s t e r n E u r op e N o r t h e r n E u r op e S ou t h e r n E u r op e W e s t e r n E u r op e E u r op e A u s t r a li aa nd N e w Z ea l a nd O cea n i a ( e x c l . AU S + N Z ) O cea n i a ( i n c l . AU S + N Z ) A l g e r i a B u r und i A ngo l a B o t s w a n a B e n i n B u r und i A l g e r i a A n ti gu aa nd B a r bud a B e li ze A r g e n ti n a A n ti gu aa nd B a r bud a C a n a d a A n ti gu aa nd B a r bud a K aza kh s t a n C h i n a B r un e i D a r u ss a l a m A f gh a n i s t a n A r m e n i a K aza kh s t a n B e l a r u s D e n m a r k A l b a n i a A u s t r i a B e l a r u s A u s t r a li a F iji A u s t r a li a E gyp t , A r a b R e p . C o m o r o s C a m e r oon L e s o t ho B u r k i n a F a s o C o m o r o s E gyp t , A r a b R e p . B a h a m a s , T h e C o s t a R i ca B o li v i a B a h a m a s , T h e G r ee n l a nd B a h a m a s , T h e K y r gy z R e pub li c K o r ea , D e m . P e op l e ’ s R e p . C a m bod i a B a ng l a d e s h A ze r b a ij a n K y r gy z R e pub li c B u l g a r i a E s t on i a B o s n i aa nd H e r ze gov i n a B e l g i u m B u l g a r i a N e w Z ea l a nd P a pu a N e wG u i n ea N e w Z ea l a nd M o r o cc o D ji bou ti C e n t r a l A fr i ca n R e pub li c N a m i b i a C o t e d ’I vo i r e D ji bou ti M o r o cc o B a r b a do s E l S a l v a do r B r az il B a r b a do s U n it e d S t a t e s B a r b a do s T a ji k i s t a n J a p a n I ndon e s i a B hu t a n B a h r a i n T a ji k i s t a n C zec h R e pub li c F i n l a nd C r o a ti a F r a n ce C zec h R e pub li c S o l o m on I s l a nd s F iji T un i s i a E r it r ea C h a d S ou t h A fr i ca G a m b i a , T h e E r it r ea T un i s i a C ub a G u a t e m a l a C h il e C ub a C ub a T u r k m e n i s t a n M ongo li a L a o P D R I nd i a C yp r u s T u r k m e n i s t a n H ung a r y I ce l a nd G r eece G e r m a ny H ung a r y V a nu a t u P a pu a N e wG u i n ea E t h i op i a C ongo , R e p . G h a n a E t h i op i a B u r und i D o m i n i ca H ondu r a s C o l o m b i a D o m i n i ca D o m i n i ca U z b e k i s t a n M a l a y s i a Ir a n , I s l a m i c R e p . G e o r g i a U z b e k i s t a n P o l a nd Ir e l a nd I t a l y L i ec h t e n s t e i n P o l a nd M i c r on e s i a , F e d . S t s . S o l o m on I s l a nd s K e ny a C ongo , D e m . R e p . G u i n ea - B i ss a u K e ny a C o m o r o s G r e n a d a M e x i c o E c u a do r G r e n a d a G r e n a d a M y a n m a r M a l d i v e s Ir a q C h i n a M o l dov a L a t v i a M a lt a L ux e m bou r g M o l dov a P a l a u V a nu a t u M a d a g a s ca r E qu a t o r i a l G u i n ea L i b e r i a M a d a g a s ca r D ji bou ti H a iti N i ca r a gu a G uy a n a H a iti H a iti P h ili pp i n e s N e p a l I s r ae l K o r ea , D e m . P e op l e ’ s R e p . R o m a n i a L it hu a n i a M on t e n e g r o N e t h e r l a nd s R o m a n i a K i r i b a ti M i c r on e s i a , F e d . S t s . M a l a w i G a bon M a li M a l a w i E r it r ea J a m a i ca P a n a m a P a r a gu a y J a m a i ca J a m a i ca S i ng a po r e P a k i s t a n J o r d a n J a p a n R u ss i a n F e d e r a ti on N o r w a y P o r t ug a l S w it ze r l a nd R u ss i a n F e d e r a ti on S a m o a P a l a u M a u r iti u s S a o T o m ea nd P r i n c i p e M a u r it a n i a M a u r iti u s E t h i op i a P u e r t o R i c o P e r u P u e r t o R i c o P u e r t o R i c o T h a il a nd S r i L a nk a K u w a it M ongo li a S l ov a k R e pub li c S w e d e n S e r b i a S l ov a k R e pub li c T ong a K i r i b a ti M o za m b i qu e N i g e r M o za m b i qu e K e ny a T r i n i d a d a nd T ob a go S u r i n a m e T r i n i d a d a nd T ob a go T r i n i d a d a nd T ob a go T i m o r- L e s t e L e b a non B r un e i D a r u ss a l a m U k r a i n e U n it e d K i ngdo m S l ov e n i a U k r a i n e T uv a l u S a m o a R w a nd a N i g e r i a R w a nd a M a d a g a s ca r U r ugu a y B e li ze B e li ze V i e t n a m O m a n C a m bod i a S p a i n D e n m a r k T ong a S e y c h e ll e s S e n e g a l S e y c h e ll e s M a l a w i V e n ez u e l a , RBC o s t a R i ca C o s t a R i ca Q a t a rI ndon e s i a E s t on i a T uv a l u S o m a li a S i e rr a L e on e S o m a li a M a u r iti u s E l S a l v a do r E l S a l v a do r S a ud i A r a b i a L a o P D R F i n l a nd S ou t h S ud a n T ogo S ou t h S ud a n M o za m b i qu e G u a t e m a l a G u a t e m a l a S y r i a n A r a b R e pub li c M a l a y s i a I ce l a nd U g a nd a U g a nd a R w a nd a H ondu r a s H ondu r a s T u r k e y M y a n m a rIr e l a nd T a n za n i a T a n za n i a S e y c h e ll e s M e x i c o M e x i c o U n it e d A r a b E m i r a t e s P h ili pp i n e s L a t v i a Z a m b i a Z a m b i a S o m a li a N i ca r a gu a N i ca r a gu a Y e m e n , R e p . S i ng a po r e L it hu a n i a Z i m b a b w e Z i m b a b w e S ou t h S ud a n P a n a m a P a n a m a T h a il a nd N o r w a y A ngo l a U g a nd a A r g e n ti n a A r g e n ti n a T i m o r- L e s t e S w e d e n C a m e r oon T a n za n i a B o li v i a B o li v i a V i e t n a m U n it e d K i ngdo m C e n t r a l A fr i ca n R e pub li c Z a m b i a B r az il B r az il A f gh a n i s t a n A l b a n i a C h a d Z i m b a b w e C h il e C h il e B a ng l a d e s h B o s n i aa nd H e r ze gov i n a C ongo , R e p . A ngo l a C o l o m b i a C o l o m b i a B hu t a n C r o a ti a C ongo , D e m . R e p . C a m e r oon E c u a do r E c u a do rI nd i a G r eece E qu a t o r i a l G u i n ea C e n t r a l A fr i ca n R e pub li c G uy a n a G uy a n a Ir a n , I s l a m i c R e p . I t a l y G a bon C h a d P a r a gu a y P a r a gu a y M a l d i v e s M a lt a S a o T o m ea nd P r i n c i p e C ongo , R e p . P e r u P e r u N e p a l M on t e n e g r o B o t s w a n a C ongo , D e m . R e p . S u r i n a m e S u r i n a m e P a k i s t a n P o r t ug a l L e s o t ho E qu a t o r i a l G u i n ea U r ugu a y U r ugu a y S r i L a nk a S e r b i a N a m i b i a G a bon V e n ez u e l a , RB V e n ez u e l a , RB A r m e n i a S l ov e n i a S ou t h A fr i ca S a o T o m ea nd P r i n c i p e C a n a d a A ze r b a ij a n S p a i n B e n i n B o t s w a n a G r ee n l a nd B a h r a i n A u s t r i a B u r k i n a F a s o L e s o t ho U n it e d S t a t e s C yp r u s B e l g i u m C o t e d ’I vo i r e N a m i b i a G e o r g i a F r a n ce G a m b i a , T h e S ou t h A fr i ca Ir a q G e r m a ny G h a n a B e n i n I s r ae l L i ec h t e n s t e i n G u i n ea - B i ss a u B u r k i n a F a s o J o r d a n L ux e m bou r g L i b e r i a C o t e d ’I vo i r e K u w a it N e t h e r l a nd s M a li G a m b i a , T h e L e b a non S w it ze r l a nd M a u r it a n i a G h a n a O m a n N i g e r G u i n ea - B i ss a u Q a t a r N i g e r i a L i b e r i a S a ud i A r a b i a S e n e g a l M a li S y r i a n A r a b R e pub li c S i e rr a L e on e M a u r it a n i a T u r k e y T ogo N i g e r U n it e d A r a b E m i r a t e s N i g e r i a Y e m e n , R e p . S e n e g a l S i e rr a L e on e T ogo W o r l d c on t a i n s a llli s t e d c oun t r i e s . GlobalNorth GlobalSouth LDC LLDC SIDS G20 EmergingMarkets OPEC LowIncome LowermiddleIncome UppermiddleIncome HighIncomeAlbania Fiji Yemen,Rep. Afghanistan AntiguaandBarbuda Australia Bangladesh Algeria Afghanistan Angola Albania AntiguaandBarbudaAustria Micronesia,Fed.Sts. Afghanistan Armenia Bahamas,The Canada Egypt,ArabRep. Angola Benin Bangladesh Algeria AustraliaBelarus Tonga Burundi Azerbaijan Barbados SaudiArabia Indonesia EquatorialGuinea BurkinaFaso Bhutan Argentina AustriaBelgium Vanuatu Angola Bhutan Belize UnitedStates Iran,IslamicRep. Gabon Burundi Bolivia Armenia Bahamas,TheBosniaandHerzegovina Tuvalu Benin Bolivia Comoros India Mexico Iran,IslamicRep. CentralAfricanRepublic Cambodia Azerbaijan BahrainBulgaria SolomonIslands Mozambique Botswana Cuba RussianFederation Nigeria Iraq Chad Cameroon Belarus BarbadosCroatia Samoa BurkinaFaso BurkinaFaso Dominica SouthAfrica Pakistan Kuwait Congo,Dem.Rep. Comoros Belize BelgiumCyprus PapuaNewGuinea Niger Burundi DominicanRepublic Turkey Philippines Libya Eritrea Congo,Rep. BosniaandHerzegovina CanadaCzechRepublic Palau CentralAfricanRepublic CentralAfricanRepublic Fiji Argentina Turkey Nigeria Ethiopia Coted’Ivoire Botswana ChileDenmark Kiribati Chad Chad Grenada Brazil Korea,Dem.People’sRep. SaudiArabia Gambia Djibouti Brazil CroatiaEstonia Bangladesh Lesotho Ethiopia Guinea-Bissau Mexico Vietnam UnitedArabEmirates Guinea Egypt,ArabRep. Bulgaria MaltaFinland Bhutan Liberia Kazakhstan Guyana France Brazil Congo,Dem.Rep. Guinea-Bissau ElSalvador China CyprusFrance Cambodia Congo,Dem.Rep. KyrgyzRepublic Haiti Germany RussianFederation Venezuela,RB Haiti Ghana Colombia CzechRepublicGreece China Djibouti LaoPDR Jamaica Italy India Liberia Honduras CostaRica DenmarkGermany India Togo Lesotho Kiribati UnitedKingdom China Madagascar India Cuba EstoniaGreenland Indonesia EquatorialGuinea Malawi Maldives China SouthAfrica Malawi Indonesia Dominica FinlandHungary LaoPDR Eritrea Mali Mauritius Indonesia Mali Kenya DominicanRepublic FranceIceland Malaysia Ethiopia Moldova Palau Japan Mozambique Kiribati Ecuador GermanyIreland Myanmar Gambia Mongolia PapuaNewGuinea Korea,Dem.People’sRep. Nepal KyrgyzRepublic EquatorialGuinea GreeceItaly Mongolia Madagascar Nepal PuertoRico Niger LaoPDR Fiji GreenlandLatvia Nepal Malawi Niger Samoa Rwanda Lesotho Gabon HungaryLiechtenstein Pakistan Mali Paraguay SaoTomeandPrincipe SierraLeone Mauritania Georgia IcelandLithuania Philippines Rwanda Rwanda Seychelles Somalia Moldova Grenada IrelandLuxembourg SriLanka Senegal SouthSudan Singapore SouthSudan Mongolia Guatemala IsraelMalta Thailand SierraLeone Tajikistan SolomonIslands SyrianArabRepublic Morocco Guyana ItalyMontenegro Timor-Leste Mauritania Turkmenistan Suriname Tajikistan Myanmar Iran,IslamicRep. JapanNetherlands Vietnam Guinea-Bissau Uganda Timor-Leste Tanzania Nicaragua Iraq Korea,Dem.People’sRep.Norway Maldives Guinea Uzbekistan Tuvalu Togo Nigeria Jamaica KuwaitPoland Grenada Comoros Zambia Vanuatu Uganda Pakistan Jordan LatviaPortugal Dominica SaoTomeandPrincipe Zimbabwe Yemen,Rep. PapuaNewGuinea Kazakhstan LiechtensteinRomania Barbados Zambia Philippines Lebanon LithuaniaSerbia AntiguaandBarbuda Uganda SaoTomeandPrincipe Libya LuxembourgSlovakia Cuba Tanzania Senegal Malaysia NetherlandsSlovenia Bahamas,The SouthSudan SolomonIslands Maldives NewZealandSpain PuertoRico Sudan Sudan Mauritius NorwaySweden Jamaica Bhutan Timor-Leste Mexico OmanSwitzerland Algeria Cambodia Tunisia Montenegro PalauUkraine Angola Bangladesh Ukraine Namibia PanamaUnitedKingdom Benin Haiti Uzbekistan Paraguay PolandCanada Botswana Kiribati Vanuatu Peru PortugalUnitedStates BurkinaFaso LaoPDR Vietnam Romania PuertoRicoAzerbaijan Cameroon Myanmar Zambia RussianFederation QatarGeorgia CentralAfricanRepublic Nepal Zimbabwe Samoa SaudiArabiaIsrael Chad Vanuatu Serbia SeychellesRussianFederation Congo Tuvalu SouthAfrica SingaporeTurkey Coted’Ivoire SolomonIslands SriLanka SlovakRepublicAustralia Congo,Dem.Rep. Timor-Leste Suriname SloveniaNewZealand Djibouti Thailand SpainKorea,Dem.People’sRep. Egypt,ArabRep. Tonga SwedenJapan EquatorialGuinea Turkey SwitzerlandSingapore Eritrea Turkmenistan TrinidadandTobagoEthiopia Tuvalu UnitedArabEmiratesGabon Venezuela,RB UnitedKingdomGambia,The UnitedStatesGhana UruguayKenyaLesothoLiberiaLibyaMadagascarMalawiMaliMoroccoMozambiqueNamibiaNigerNigeriaRwandaSenegalSierraLeoneSomaliaSouthAfricaSouthSudanSudanSyrianArabRepublicTogoTunisiaUgandaTanzaniaZambiaZimbabweSeychellesSaoTomeandPrincipeMauritiusMauritaniaGuinea-BissauGuineaComorosBurundiBelizeBahamas,TheArgentinaBoliviaBrazilChileColombiaCostaRicaCubaDominicanRepublicEcuadorElSalvadorGuatemalaHaitiHondurasJamaicaMexicoPanamaParaguayPeruPuertoRicoSurinameTrinidadandTobagoUruguayVenezuela,RBNicaraguaGuyanaGrenadaDominicaBarbadosAntiguaandBarbudaIraqAfghanistanArmeniaBahrainIran,IslamicRep.JordanKazakhstanKuwaitKyrgyzRepublicLebanonOmanQatarSaudiArabiaTajikistanTurkmenistanUnitedArabEmiratesUzbekistanYemen,Rep.