The bifurcation of inequality trends in SSA over

Transcription

1 The bifurcation of inequality trends in SSA over Giovanni Andrea Cornia (with Bruno Martorano e Chiara Gualdani) University of Florence WIDER 30th Anniversary Conference, Helsinki,

2 Structure of presentation 1. Building Gini database IID-SSA and bifurcating Gini trends 2. Explaining inequality changes over (i) Changes in output structure btw sectors inequality (ii) Changes in within-sector ineq. (for agriculture, urban sector) (iii) Population growth and dependency rates (iv) Macro policy variables (inflation, REER) (v) Redistributive policies (tax, transfers, education, health) (vi) External factors (tot, FDI, remittances, aid, HIPC, tariffs) (vii) Shocks (cellphones, AIDS, wars) (viii) Governance indicators 3. Macro panel regression of determinants of inequality 4. Tentative conclusions on drivers of inequality

3 1.Building IID-SSA database Gini databases: WIIDv3, SWIID, POVCAL, WYD, I2D2, (RIGA), nat. data compiled IIS-SSA Gini database for countries with at least 4-5 well spaced observations from above datasets (excluding SWIID) btw and 2011 Select best data (fully documented) from above sources through protocol Eliminate 6-7 outliers (e.g. 20 pts Gini shifts from a year to the next) Retained 29 countries (on 48) which represent > 90% SSA s pop and GDP About 240 observed and controlled Gini for out of 580 cells To complete Gini matrix, we interpolate linearly 340 data btw observed Gini s Grouped 29 countries by their Gini trends into 4 groups: rising, falling, U and

4 1. Falling trend in mean Gini for 29 SSA countries, unweighted (top) and weighted (bottom).but

5 1. At disaggregated level bifurcation in (unweighted) Gini trends:

6 2.(i) Growth & Gini no systematic relation y = x falling inequality R 2 = y = x rising inequality R 2 =

7 2(i) Growth pattern sub-optimal transition output shares - fall in (mostly) low-ineq-productivity agriculture (left scale)) - Slow rise-stagnation-fall in low-ineq manufacturing - Rises in oil-mining sector (right scale) due to world prices +discoveries - Rises in formal (FIRE, pub adm, trade) & informal urban services Trend in v.a added share of main sectors Source: UNCTADSTAT

8 2.(i) Rise in production & exports of mining-oil resources Country Country Country (a) share of GDP > 20% (b) share of GDP btw % (c) share of GDP btw 5-10% Angola B.Faso C.Ivoire Chad Burundi Ethiopia Congo DR Camerun Ghana Congo Rep G Bissau Malawi Eq. Guinea Guinea Mozamb Gabon Liberia S.Leone Mauritania Mali Tanzania Nigeria S.Africa Uganda Zambia Sudan Zimbab Average (rising) Average (rising) Average constant At least 18 countries depend on (un-equalizing) oil/min rents

9 2(i) Sub-optimal struct. transition from low- to high- Gini sectors re-primarization, informalization, premature tertiarization Gini VA shares(x) & Gini coefficients (y) 29 countries, (UNCTAD data VA Agriculture gini Trade, restaurants and hotels VA share of manufacturing stagnated or declined, incl. due to trade liberalization Construction VA Manufacturing

10 Finance, insurance, real estate and business services Gini Mining Community, social and personal services Government services

11 2(i) Slow urbanization (often unequalizing). With no change in structural transformation pattern, rising Gini in future Region and country Less Developed Regions Change over prior year SSA Change over prior year Eastern Africa Burundi (min) Ethiopia Middle Africa Chad Southern Africa South Africa Western Africa Niger Greater focus on rising land yields, manufacturing, construction is needed, à la Ranis Fei

12 2 (ii) equalizing changes in within sector inequality: rising output/c in agriculture?

13 2(ii) at times rise in food output driven by rising yields

14 (ii) Gini change depends on land distribution Stronger effect of agricultural output rise where land distribution is more egalitarian What happened to land distribution? State and local level land titling programs Land reform confirmed (Ethiopia) &certification (Rwanda) Rural pop. growth may lead to land concentration (Niger) Land grabs? Potentially un-equalizing

15 Population dividend or inequality time bomb? 2.(iii) Demography: little change + Gini Trends in population growth rates in SSA Relation btw pop growth (or dep rate)&gini WORLD Less dev.ped regions SSA East Africa West Africa gini No relation btw pop growth and Gini. Rwanda Rwanda Rwanda Rwanda Sierra Leone Sierra Leone Sierra Leone Sierra Leone Sierra Leone Rwanda Rwanda Guinea Guinea Guinea Guinea Rwanda Rwanda R Rw Niger Population growth (annual %) But micro data tell a different story

16 (iv) Macro policies: inflation control +RER - Budget deficits: falling and little related to Gini - REER: push to appreciation in 2000s in resource-cfa cties - Rising trade liberalization but fall in manufacturing/gdp Inflation and Gini % change in CPI (x) and Gini(y) Inflation, consumer prices (annual %)

17 2.(i) Malawi: average tariff rate (left scale) & manufact. v.a. share (right scale) WDI data

18 2 (V) Education: rapid increase in 1ary (good for cognitive purposes). Less for 2ary (technical knowledge). Affect skill premium &yrs of educ. Relation btw average yrs of educ (x) & Gini educ (y) Educational inequality of L.F. 6 to 9 Average n. of yrs of educ of l.f.

19 2(v) Social spending became better targeted, but still regressive as share of transfer - but not as a share of income of recipients Source: Ferreira (2014 Share of bottom (blue)& top quintile (green) of yrs who completed grade 6, late 2000s

20 2.(v) Aggregate link btw social services (e-h- ss) & Gini - but micro evidence is weak With MDGs &HIPC, some drive to raise social spending but modest effects so far Social spending (right scale) and Gini (left)

21 2(v)TOT, remittances, FDI, Aid, HIPC: favorable Distributive effects: ok for non-mining tot, remittances, HIPC; FDI?

22 2(vi) Shocks: wars & conflicts decline

23 2(vi). Shocks: HIV rate trend declined from mid 2000s - & relation btw HIV (x-axis) & Gini (y-axis) Trend in HIV incidence HIV incidence and Gini

24 2.(vi) Technological shocks: Mobile & internet: greater market integration? Gini Mobile phones Internet users Mobile cellular subscriptions (per 100 people)

25 2.(vii) Governance: Trends in types of regimes, But these & other indexes (FHI, Polity 2) unable to capture changes in real governance, political participation, corruption, etc.

26 3. Econometric test on macro panel Test based on fairly balanced panel of 29 countries & 20 yrs Use both immediate and underlying determinants of inequality A few variables ( tariff rate) excluded for lack of data, or as uncorrelated Some variables (e.g. dependency ratio) not significant on macrodata but significant on micro data LSDV estimator - Issues of persistence/endogeneity being addressed 80 % of the hypotheses discussed above are verified: encouraging More work to do explore interactions for SSA subgroups needed

27 Regression results: LSDV estimates; dep. variable: family consumption/c

28 4. Tentative conclusions: drivers of inequality bifurcation (B) Bifurcation due to different endowments, changes in output structure, policies Immediate causes of inequality (i) between sectors inequality. In several countries, sub-optimal structural transition to capital/skills-intensive sectors (mining, FIRE, commerce, pub. admin), leading to re-primarization, informalization, premat.tertiarization, & + between-sectors inequality GDP may rise but inequality grows & poverty allev. elasticity of growth (PAEG) falls Some countries shifted to high-yielding labor-intensive agric, manufacturing, construction & labour intensive services (RNAA). Increase in rural 2ary education is equalizing (accelerates use of modern farming techniques) but reverse causation also true

29 4. Continued (ii) within sector ineq: distribution of land and capital did not change, that of 2ary education worsened. tenancy reforms improved security/incentives of tillers land grabs, commercialization & pop. growth rise in furute land Gini This may not depress output but will affect rural inequality & PAEG. Human capital: despite strong gains at 1ary level, 2ary educ rose slowly, is poorly distributed, & contributes to inequality via rise in skill premium (iii) population, no slow down (exc.southern Africa) leads to inequalizing pressure on land, distress urbanization, dependency rate, wage rate Future population dividend or population disaster?

30 4. Continued (iv) Macro policy (RER and CPI) affect inequality (v) redistribution. Revenue/GDP ratio rose taxation is progressive in Ethiopia etc but not in mineral enclaves Weak equalization due to social spending as weak institutions limit redistribution, esp. in countries w/ large rents 123 pilot cash transfers. Where properly designed and embedded (S.Africa & Ethiopia) larger CT reduce nationwide Gini Main task now is to develop redistributive institutions

31 4. Continued underlying causes of inequality (vi) Changes in external conditions. Gains in TOT equalizing but not in mining/oil countries remittances equalizing (against most theories) FDI (mostly in resource sector) non significant or diseq Aid non significative but debt cancellation (HIPC) progressive (vii) Shocks Fall in HIV/AIDS & war intensity were equalizing Diffusion of ITC (techno shock) not significant or disequalizing (viii) Democracy & governance Found not significant effects but democracy difficult to proxy