PPP-based Stratification of CIS EU/OECD Economies

Interstate Statistical Committee of the Commonwealth of Independent States (CIS-Stat) PPP-based Stratification of CIS EU/OECD Economies Andrey E. Kosarev IARIW / HSE conference Experiences and Future Challenges in Measuring Income and Wealth in CIS Countries and Eastern Europe Moscow 17-18 September 2019

Outline I. Stratification II. Purchasing Power Parity III. Moving Linear Segment procedure IV. Harrod-Balassa-Samuelson stratification V. Engel s law stratification VI. Conclusions 2

Stratification Economic policy measures inducing a GDP increase, evoke : Resulting increase of GDPpc provokes changes of any other indicator in a specific proportion This proportion is assessed separately within each group of similar economies Groups of economies are defined by similarity of tendencies in changes provoked by GDPpc tendency-groups 160 PLI, % 140 120 100 80 Expected changes of an endogenous indicator depend on the tendency-group 60 40 20 0 0 20,000 40,000 60,000 80,000 100,000 120,000 GDPpc, USD 3

Purchasing Power Parity Purchasing Power Parity (PPP) ratio between currencies, which takes into account the price levels in different economies. PPPs are computed within a joint work of countries, a largest international project : Global International Comparison Program (ICP), implemented by the decision of the UN Statistical Commission. 2014 ICP : PPPs for a joint group of 54 countries, CIS EU/OECD. PPP-based indicators allow to detect tendency groups and to substantiate the stratification within the set of 54 CIS EU/OECD countries. A special Moving Linear Segment procedure (MLS-procedure) was elaborated for these purposes. 4

Moving Linear Segment (1/5) 160 PLI, % 140 120 100 80 60 40 20 0 0 20,000 40,000 60,000 80,000 100,000 120,000 GDPpc, USD Countries are ranked following the ascending GDP per capita in PPP terms. 5

Moving Linear Segment (2/5) PLI, % 160 140 120 100 80 60 40 20 0 0 20,000 40,000 60,000 80,000 100,000 120,000 GDPpc, USD For each group of consequent points a linear function indicator under analysis = a0 + a1 * GDPpc is estimated e.g.: for a set of 54 countries 36 linear segments, each including 19 points, are estimated 6

Moving Linear Segment (3/5) PLI, % 160 140 120 100 80 60 40 20 0 0 20,000 40,000 60,000 80,000 100,000 120,000 GDPpc, USD Structural break points have to be detected 7

Moving Linear Segment (4/5) Structural break point: sbc(j) = ( a(j+1) + + a(j+k)) / (a(j-1) + + a(j-k) ) j = (k+1), , (J-k) sbc(j) structural break coefficient at the point j a(j) linear regression angle for the segment with the center at the point j K number of linear regression segments to the right and to the left of the point j J total number of segments If a local extremum sbc(t) exists, then the point t is a break point dividing two neighboring homogeneous tendency-groups . Analysis of CIS EU/OECD 54-points set: 36 19-points segments (J=36), K=3 8

MLS (5/5) HBS stratification (1/2) 160 PLI, % 140 120 100 80 60 40 20 0 0 20,000 40,000 60,000 80,000 100,000 120,000 GDPpc, USD The moving linear segment structural break points define the groups each having a different tendency of PLI changes depending the GDPpc. 9

HBS stratification (2/2) PLI = PPP / XR the eXchange Rate Price Level Index is the ratio of the Purchasing Power Parity to Harrod-Balassa-Samuelson effect: PLI is positively correlated with GDP per capita. 54 CIS EU/OECD economies, in PPP 2014 terms, three tendency-groups can be considered: GDPpc Increase of PLI (%) accompanying the USD 1000 increase of GDPpc tendency characteristics USD, PPP 2014 t R < 28500 1.6 8.99 86.6% 28500 41500 3.1 9.61 89.5% > 41500 0.2 1.23 29.3% 10

Engels law stratification (1/3) Engel s law: as income rises, the proportion of income spent on food falls. Several indicators from PPP publications can be considered in this context, e.g.: - ratio (1) expenses on clothing & footwear to expenses on food & non- alcoholic beverages increases as GDP per capita increases ; - ratio (2) expenses on household furnishings, equipment and maintenance to expenses on food & non-alcoholic beverages increases as GDP per capita increases 11

Engels law stratification (2/3) 0.9 Cl.&Foot. / Food Ratio (1) expenses on 0.8 clothing & footwear to 0.7 0.6 expenses on food & 0.5 non-alcoholic 0.4 0.3 beverages increases as 0.2 GDP per capita increases 0.1 0.0 0 40,000 80,000 120,000 GDPpc, USD GDPpc Increase of Ratio 1 accompanying the USD 1000 increase of GDPpc tendency characteristics USD, PPP 2014 t R < 28000 0.007 3.88 61.3% 28000 34000 0.018 4.03 68.9% > 34000 0.005 2.24 42.3% 12

Engels law stratification (3/3) 0.9 FEM / Food Ratio (2) expenses on household furnishings, equipment and maintenance to expenses on food & non-alcoholic beverages increases as GDP per capita increases 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0.0 0 40,000 80,000 120,000 GDPpc, USD GDPpc Increase of Ratio 2 accompanying the USD 1000 increase of GDPpc tendency characteristics USD, PPP 2014 t R < 29000 0.008 3.19 53.0% 29000 45000 0.018 6.36 79.2% > 45000 0.003 1.79 40.9% 13

Conclusions (1/2) 2.5 Clothing & Footwear / Food & Non-alc.beverages 2.0 denominated by average mean Household furnishings, equipment, maintenance / Food & Non-alc.beverages 1.5 1.0 PLI 0.5 0.0 0 20,000 40,000 60,000 GDPpc 80,000 100,000 120,000 14

Conclusions (2/2) PPP-based stratification: in PPP 2014 terms, economies within a joint CIS EU/OECD set can be divided into 3 income groups - low-income economies having approx. GDPpc < USD 28500 - middle-income economies having approx. USD 28500 < GDPpc < USD 40000 - high-income economies having approx. GDPpc > USD 40000 Each group has its specific tendency to respond to economic growth. While elaborating economic policy measures, it is important to take into account that expected effects would depend on what group the economy under consideration belongs to. 15

Thank you! 16