From the SelectedWorks of Prof. Dr. Mohammad Reza Farzanegan September, 2012 Demographic transition in resource rich countries: a bonus or a curse? Kjetil Bjorvatn Mohammad Reza Farzanegan Available at: https://works.bepress.com/mr_farzanegan/15/
Demographic Transition in Resource Rich Countries: A Blessing or a Curse? Kjetil Bjorvatn (NHH, Norway) and Mohammad Reza Farzanegan (CNMS, Germany) farzanegan@staff.uni-marburg.de www.uni-marburg.de/cnms/wirtschaft Jahrestagung des Vereins für Socialpolitik 2012, Georg-August-Universität Göttingen
Agenda 1. Motivation 2. Contribution of this study 3. Data and Empirical Approach 4. Results 5. Summary and conclusion 2
Motivation The Middle East and North Africa (MENA) is home to the most dramatic demographic transition in the world today, with its working-age population rising at a rate of almost three percent per year. In other regions, similar demographic changes have constituted a powerful source of economic growth (e.g., East Asia: 1/3 of growth!). In contrast, the increasingly young populations in MENA appear to constitute a force of social and political unrest rather than economic progress 3
Motivation: Same working age, different results (1976-2006) Region GDP Share of Economic Democratic Resourc per capita (percent growth) working age (percent growth) IQ IQ e rents (percent of GDP) East 670 6.70 071 0.71 051 0.51 045 0.45 779 7.79 Asia MENA 0.79 0.69 0.51 0.15 24.34 Latin A 1.00 0.53 0.47 0.70 5.21 World 1.55 0.38 0.55 0.54 8.26 4 Source: World Bank (2011) and authors calculations
Motivation This study argues that resource wealth is a key factor for understanding income effects of demographic transition. We develop a simple model highlighting g g how the income effect of increased labor supply may depend on resource rents. Using panel data covering the period from 1982 2006 2006 for more than 120 countries, we find a negative interaction effect between resource wealth and demographic transition on national income. 5
Contribution Our paper adds to the traditional Dutch disease argument by showing that: resource rents not only affect the structure of the labor market and economic growth per se,, but also reduce the capacity of the economy to productively absorb increases in the work force. We demonstrate that resource rents are an important conditioning factor for the growth effect of the demographic transition, even when controlling for institutions. 6
Data and Empirical Approach Panel regressions for more than 120 countries from 1982-2006, using five years average We estimate following model: INCOME k it, jx jit,, 1DEMOGit, 2RES it, + 3( DEMOGit, RES it, ) + i t it, j 11 INCOME: log of real GDP per capita DEMOG: share of working age population in total population RES: share of total resource rents in GDP (or per capita) X: controls such as education, investment, inflation, government spending, trade openness, life expectancy, institutions and their interactions with DEMOG and RES 7
Data and Empirical Approach k INCOME it, jx jit,, 1DEMOGit, 2RES it, + 3( DEMOGit, RES it, ) + i t it, j 1 The effect of a change in DEMOG on INCOME is ( INCOME ) ( DEMOG) 1 3RES The model predicts that: the sign of this interaction term should be negative the higher the share of natural resource rents in the economy, the lower would be the income effect of a demographic transition 8
Results Our baseline specification confirms the theoretical predications: Income effects of demographic transition is moderated significantly by the natural resource wealth. Countries with high natural resource wealth may indeed face a demographic curse instead of demographic dividend (or bonus). Robustness tests: Results hold for alternative natural resource wealth proxy Results hold for alternative demographic transition proxy Results hold using robust estimator t due to possible outliers effects Results hold using GMM and 2SLS, accounting for possible endogeneity of DEMOG, RES and their interaction Results hold using 5 years lag of RHS variable Results hold using annual data instead of 5 years average of data Results hold across different geographical and income samples 9
Results Country y and time fixed effects panel regressions (1982-2006)(5 years average, 5 years lag of RHV)); using ICRG Dep.: Logarithm of real GDP per capita M1 M2 M3 M4 DEMOG 1.882*** 1.352** 0.94 0.90 (3.23) (2.52) (1.28) (0.97) RES 2.53*** 2.19** 2.45*** 236 2.36 (3.21) (2.61) (2.76) (1.58) DEMOG* RES -0.639*** (-3.31) 31) -0.55*** (-2.65) -0.63*** (-2.78) -0.60 (-1.62) All &ICRG, All &ICRG, ICRG^2, All ICRG^2, RES*ICRG, Controls All &ICRG, DEMOG* ICRG^2 ICRG & RES*ICRG, DEMOG* ICRG obs. 479 308 308 308 Countries 152 123 123 123 ( 162) -0.043 043 (-0.06) RES*DEMOG* ICRG 10 R-sq. 0.59 0.65 0.65 0.65
Results Country and time fixed effects panel regressions (1982-2006) (5 years average, 5 years lag of RHV)); using POLITY DEMOG RES DEMOG* RES Controls M5 M6 M7 1.80*** 1.24* 0.71 (2.63) (1.79) (0.80) 2.22** 198** 1.98 073 0.73 (2.39) (2.00) (0.46) -0.56** -0.50** -0.19 (-2.47) (-2.02) (-0.48) All &Polity, All &Polity, Polity^2, Polity^2, All & RES*Polity, RES*Polity Polity, DEMOG* & Polity^2 Polity & DEMOG* RES*DEMOG Polity *Polity obs. 423 423 423-0.567 (-1.00) Countries 135 135 135 11 R-sq. 0.60 0.61 0.61
Results Marginal income impacts of demographic transition (based on M1) (Log of INCOME) (log of DEMOG) 188. 063. (log of RES) In countries with a higher than (log) RES of 2.98 (1.88/0.63), demographic transition leads to lower per capita income, i.e., a demographic curse. 12
Results: Countries in the demographic curse zone (2002-2006) Algeria, Angola, Azerbaijan, Bahrain, Bolivia, Brunei, Chad, Congo, Dem. Rep., Congo R., Ecuador, Equatorial Guinea, Gabon, Iran, Iraq, Kazakhstan, Kuwait, Libya, Nigeria, Oman, Papua New Guinea, Qatar, Russia, Saudi Arabia, Syrian Arab Republic, Trinidad and Tobago, Turkmenistan, UAE, Uzbekistan, Venezuela, Yemen 13
Summary and Conclusion Why some countries benefit from their working age population, while others, mainly MENA countries, not? We study the mechanism through which the demographic transition feeds economic development, focusing on the role of resource rents in this process. In our theoretical model we show that a larger working age population has the potential of stimulating economic development, but also that resource rents may dampen this income stimulus, potentially even reversing it. Panel data regression for more than 120 countries from 1982-2006 confirms our expectations. 14