Economics growth and the middle class in an economy in transitions the case of russia

Economic Studies in Inequality, Social Exclusion and Well-Being Series Editor: Jacques Silber Zoya Nissanov Economic Growth and the Middle Class in an Economy in Transition The Case of Russia www.ebook3000.com Economic Studies in Inequality, Social Exclusion and Well-Being Series editor Jacques Silber, Ramat Gan, Israel More information about this series at http://www.springer.com/series/7140 www.ebook3000.com Zoya Nissanov Economic Growth and the Middle Class in an Economy in Transition The Case of Russia 123 Zoya Nissanov Department of Economics and Business Management Ariel University Ariel Israel ISSN 2364-107X ISSN 2364-1088 (electronic) Economic Studies in Inequality, Social Exclusion and Well-Being ISBN 978-3-319-51093-4 ISBN 978-3-319-51094-1 (eBook) DOI 10.1007/978-3-319-51094-1 Library of Congress Control Number: 2016962034 © Springer International Publishing AG 2017 This work is subject to copyright All rights are reserved by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed The use of general descriptive names, registered names, trademarks, service marks, etc in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations Printed on acid-free paper This Springer imprint is published by Springer Nature The registered company is Springer International Publishing AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland www.ebook3000.com Acknowledgements I would like to thank Jacques Silber for his constructive comments I also thank Raphael Franck and Grazia Pittau for their comments on earlier versions of some of the chapters Finally, special thanks to the “Russia Longitudinal Monitoring survey, RLMS-HSE”, conducted by the National Research University Higher School of Economics and ZAO “Demoscope” together with Carolina Population Center, University of North Carolina at Chapel Hill and the Institute of Sociology RAS, for making these data available v Contents What Does the Middle Class Refer To? 1.1 Importance, Measurement and Characteristics of the Middle Class 1.1.1 Importance of the Middle Class 1.1.2 Measurement of the Middle Class 1.1.3 Characteristics of the Middle Class 1.2 The Russian Middle Class 1 On the Transition in Russia 2.1 Russia in Transition 2.2 Data 2.2.1 Data description 2.2.2 Panel Dataset 2.2.3 Summary Statistics 9 11 11 13 14 Distributional Change and What Happened to the Middle Class in Russia 3.1 Review of the Literature 3.1.1 Relative Distribution 3.1.2 Decomposition 3.1.3 Relative Polarization 3.2 Empirical Results 3.2.1 Relative Polarization 3.2.2 Decomposition 3.3 Conclusions 21 22 22 23 24 25 25 29 35 Bipolarization and the Middle Class in Russia 4.1 Review of the Literature 4.1.1 Bipolarization Curves 4.1.2 Measures of Bipolarization 4.1.3 Extensions of the FW Measure 37 37 38 40 41 vii www.ebook3000.com viii Contents 4.1.4 Decomposition of the Bipolarization Measure 4.1.5 Inequality and Bipolarization 4.1.6 Bipolarization and Mobility 4.2 Empirical Results 4.2.1 Bipolarization Measures 4.2.2 Decomposition of the Bipolarization Measure (FW) by Income Sources 4.3 Conclusions 42 43 44 45 45 51 55 57 57 57 On Polarization in Russia 5.1 Review of the Literature 5.1.1 Polarization: Definition and Properties 5.1.2 Measuring Polarization with an Arbitrary Number of Poles 5.1.3 Decomposition of Polarization Indices 5.2 Empirical Results 5.2.1 Measuring Polarization 5.2.2 Decomposition of Polarization Measures by Income Sources 5.3 Conclusions 58 62 62 62 65 68 The 6.1 6.2 6.3 Socio-Economic Characteristics of the Middle Class Methodology Empirical Results Conclusions Income Mobility and the Middle Class 7.1 Methodology: Income Mobility 7.2 Empirical Results 7.2.1 Mobility Between Income Groups 7.2.2 Income Growth 7.2.3 Mobility Within Income Groups 7.3 Conclusions 71 71 74 92 93 93 94 94 98 99 101 Concluding Comments 103 Appendices 107 References 113 List of Figures Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 2.1 2.2 2.3 2.4 2.5 2.6 2.7 2.8 2.9 3.1 3.2 3.3 3.4 3.5 3.6 3.7 3.8 3.9 3.10 3.11 3.12 3.13 3.14 3.15 3.16 3.17 Figure 4.1 Figure 4.2 Average incomes (in June 1992 rubles) Equivalent incomes Ratio of mean to median Share of zero-incomes Gini index Theil entropy measure Between-group inequality Within-group inequality The share of the middle class (equivalent incomes) PDF overlays for 1992–1995 PDF overlays for 1995–1998 PDF overlays for 1998–2001 PDF overlays for 2001–2003 PDF overlays for 2003–2005 PDF overlays for 2005–2008 PDF overlays for 1992–2008 Decomposition of the relative distribution for 1992–1993 Decomposition of the relative distribution for 1993–1995 Decomposition of the relative distribution for 1995–1996 Decomposition of the relative distribution for 1996–1998 Decomposition of the relative distribution for 1998–2001 Decomposition of the relative distribution for 2001–2002 Decomposition of the relative distribution for 2002–2005 Decomposition of the relative distribution for 2005–2008 Decomposition of the relative distribution for 1992–2008 Decomposition of the relative distribution for 1992–1996 and 1996–2005 First degree bipolarization curve Second degree bipolarization curve 14 15 15 16 16 16 17 17 18 25 25 26 26 27 27 29 29 30 31 31 32 32 33 33 33 34 39 39 ix www.ebook3000.com x List of Figures Figure 4.3 Figure 4.4 Figure 4.5 Figure 4.6 Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure Figure 4.7 4.8 4.9 5.1 5.2 5.3 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 6.10 Foster-Wolfson bipolarization measure based on Lorenz curve FW-index WT-index Distance between mean and median incomes (divided by median) PG index FW-index for positive incomes Gini index for individual incomes DER-index, a = 0.5 Identification component of the DER-index DER-index for positive incomes Estimated proportions of the groups Estimated normalized incomes of the groups Age of the main earner Households where the main earner is male Households in rural area Households in Moscow and St Petersburg Households in Northern Caucasus University graduates Pensioners (retired main earners) Self-employed main earners 41 45 47 48 50 52 52 63 64 66 77 77 81 81 81 82 82 83 83 83 P P PER a; qị ẳ Kiẳ1 Kjẳ1 pi pj pai li À lj The critical aspect here is to minimize the error, which is expressed as the average of income distance within all groups An n-group representation q is equivalent to transforming the original Lorenz curve into its n-piecewise linear representation Thus, the minimization of the error is equivalent to minimizing the area between the original and piecewise Lorenz curves Hence, it can be expressed as the inequality in F minus the inequality in its representation q as measured by the Gini coefcient (G): eF; qị ẳ GðFÞ À GðqÞ That leads to the measure EGR finally proposed to account for the degree of polarization in F with PEGR F; a; bị ẳ PER a; qị bẵGFị Gqị 5:1:3ị A by-product of this extension is that it provides a connection between the EGR and FW: the FW measure is equal to the EGR for the case in which the simplified two-spike representation has been chosen to be symmetric and the relevant parameters satisfy α = β = Esteban, Gradín and Ray (2007) consider that the best representation is the one giving the highest level of polarization over all possible group representations They showed that when α = 1, EGR is generally maximized when society is divided into three groups; but for higher values of α, the two-group representation leads to slightly higher levels of extended polarization www.ebook3000.com 60 On Polarization in Russia Gradin (2000), using the EGR measure, suggested two approaches to the analysis by sub-populations: group polarization and explained polarization The first approach enables to measure polarization in a population which is divided by sub-groups according to social characteristics like race, education, occupation, region, etc In this case the within-group heterogeneity is expected to be higher than the between-groups one; therefore it is possible to find negative values in (5.1.3) The most relevant characteristics are those that show both between-groups heterogeneity and within-groups homogeneity Gradin (2000) normalized the EGR index to avoid negative values and defined Group Polarization (GP) as: GPðF; a; b; qc ị ẳ PER a; qc ị bẵeF; qc Þ À 1Š ð5:1:4Þ where qc is a collection of numbers ðq1 ; ; qn ; m1 ; ; mn Þ, qi is the share of group i in total population and mi is the average income of group i The second approach enables to find the most relevant characteristic which can explain an observed level of polarization Explained Polarization (EP) measures the degree to which a characteristic explains polarization, when groups are determined by income intervals For α = EP is equal to the ratio of between-groups inequality (when groups are determined by relevant characteristic) to total inequality in income distribution In this case EP falls in the interval [0, 1] The higher EP, the better a given characteristic explains income polarization Lasso De La Vega and Urrutia (2006) claimed that the EGR measure does not obey one of the basic features of polarization, according to which higher heterogeneity across groups should increase polarization They propose an alternative formulation for the EGR index: Pa; bị ẳ A K X K X iẳ1 jẳ1 pi pj pai ð1 À Gi Þb yi À yj ð5:1:5Þ where β is a weight assigned to the homogeneity in the identification function and Gi is the Gini inequality within group i Duclos, Esteban and Ray (2004) extended the ER measure for the case of continuous distributions described by density functions They developed a polarization measurement theory to overcome the drawback of the ER index (its discontinuity and the fact that the distribution is assumed to be bunched into a finite number of groups) For every distribution function F with density f and mean µ, they proposed the following polarization measure (DER index): R Ry Pa Fị ẳ f yịa ayịdFyị where ayị  l ỵ y2Fyị 1Þ À À1 xdFðxÞ y represents the alienation felt by an individual with income y A natural estimator of a DER measure for a random sample of n iid observations of income yi , i = 1,…,n, ^ drawn from distribution F(y) where incomes are ordered as y1 y2 yn and l is the sample mean, is given as 5.1 Review of the Literature 61 ^ ¼ nÀ1 Pa ðFÞ n X ^f ðyi Þa ^aðyi ị 5:1:6ị iẳ1 ^ ỵ yi n1 2i 1ị 1ị n1 with ^ ayi ị ẳ l iP jẳ1 ! yj ỵ yi where a ẵ0:25; and ^f yi ịa is estimated non-parametrically using kernel estimation procedures If = 0, Paẳ0 Fị is equal to the twice Gini coefficient It is possible to decompose the DER index as follows: Pa ðf Þ ẳ aia ẵ1 ỵ q 5:1:6bị where  a is the average alienation, ia is the average α-identification and q is the normalized covariance between them Zhang and Kanbur (2001) find that the polarization measures which they tested in their empirical comparisons (ER, WT, FW) not lead to very different results from the standard measures of inequality, at least in the specific context of China They proposed an alternative polarization index (KZ), which is based on the Theil index of inequality, and where polarization is defined as the ratio of the between-group inequality to the within-group inequality: PKZ ¼ Iðl1 e1 ; ; lk ek Þ PK g wg Ig ð5:1:7Þ where K refers to the number of groups, Ig to the Theil inequality index within group g, wg to the weight of group g in the total population, lg to the mean of the gth group and eg is a vector of 1’s of length ng (that is the population of the gth group) The Theil index is a specific case of the Generalized Entropy (GE) measure, when c is equal to or 1, and where GE is expressed as: n c o K P > yi > > f ðy Þ À c 6¼ 0; i > l > > i¼1 >

iẳ1 > >   > K > P > l > c¼0 : f yi ị log yi iẳ1 The similar behavior of polarization (ER and FW) and inequality measures was also reported in Fedorov (2002) for Russian regions He proposed to modify the KZ index and to define the index as the ratio of between-group inequality to total inequality for the following two reasons: (i) if within-group inequality is small, then www.ebook3000.com 62 On Polarization in Russia even small changes in within-group inequality from one period to another will lead to large swings in the results; (ii) the modified measure has an intuitive interpretation as the share of between-group inequality in the total one 5.1.3 Decomposition of Polarization Indices Araar (2008) proposed to decompose the DER index (5.1.6) by population groups as follows: Pẳ X g /g1 ỵ a w1a g Rg Pg ~ ỵP R Rg ẳ where ag yịpg yịfg yị1 ỵ a dy R /g ag yịfg yị1 ỵ a dy 5:1:8ị ~ denotes the DER index when the within-group polarization or inequality is P ignored, /g is the population share of group g, wg is the income share of group g, fg is the density function for group g, ag ðyÞ is the alienation for the individual at the level of its group g and pg ðyÞ is the local proportion of individuals belonging to group g and having income y If the groups of incomes not overlap, Rg = When α = 0, then Rg = and this decomposition is similar to that of the Gini index Araar (2008) also proposed to decompose the DER index by income sources: P¼ l1Àa Z f yị 1ỵa ayịdy ẳ X R wk k f yị1 ỵ a ak yịdy X ẳ wk CPk wak laÀ1 k ð5:1:9Þ where CPk is the pseudo-polarization index of income source k 5.2 5.2.1 Empirical Results Measuring Polarization The results of the DER index, based on (5.1.6) with α = 0.5 are given in Table 5.1 and Fig 5.1 To make the comparison easier, Duclos, Esteban and Ray (2004) suggested dividing all indices by two, so that if α = 0, the DER index is equal to the Gini coefficient Table 5.2 summarizes the main changes in the DER-index The DER increased between 1992 and 1996 and then decreased Over the whole period, the DER changed significantly only when working with individual incomes Table 5.3 gives the components of the DER index (see Eq 5.1.6b), namely the average alienation, average identification and the normalized covariance between 5.2 Empirical Results 63 Table 5.1 The DER-index (α = 0.5) Year Individuals CI 95% Households CI 95% Equivalent incomes [CI 95%] 1992 1993 1994 1995 1996 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 0.317 0.320 0.367 0.393 0.505 0.456 0.406 0.399 0.384 0.379 0.352 0.324 0.339 0.309 0.305 0.271 0.284 0.291 0.289 0.311 0.294 0.294 0.293 0.285 0.278 0.280 0.278 0.292 0.273 0.282 0.251 0.261 0.271 0.272 0.298 0.272 0.266 0.256 0.251 0.247 0.250 0.244 0.255 0.241 0.244 [0.311 [0.314 [0.358 [0.384 [0.493 [0.445 [0.397 [0.384 [0.375 [0.371 [0.345 [0.319 [0.332 [0.305 [0.300 0.322] 0.326] 0.376] 0.402] 0.517] 0.468] 0.416] 0.414] 0.393] 0.387] 0.359] 0.330] 0.345] 0.314] 0.310] [0.265 [0.274 [0.279 [0.278 [0.302 [0.283 [0.284 [0.281 [0.273 [0.271 [0.266 [0.269 [0.276 [0.267 [0.270 0.277] 0.295] 0.302] 0.299] 0.320] 0.304] 0.304] 0.305] 0.298] 0.285] 0.294] 0.288] 0.308] 0.279] 0.294] [0.245 [0.251 [0.261 [0.260 [0.289 [0.263 [0.254 [0.247 [0.242 [0.241 [0.240 [0.234 [0.243 [0.235 [0.232 0.257] 0.271] 0.282] 0.284] 0.308] 0.282] 0.278] 0.265] 0.261] 0.253] 0.261] 0.253] 0.268] 0.248] 0.256] 0.60 0.55 DER-index 0.50 0.45 0.40 0.35 0.30 0.25 0.20 0.15 1992 1993 1994 1995 1996 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 year individual incomes households equivalent incomes Fig 5.1 DER-index, a = 0.5 them The alienation component is equal to the Gini index and its plot is given in Fig 2.5 (Chap 2); the identification component is illustrated in Fig 5.2 The results show that the increase in the DER for individual incomes during the 1992–1996 period was due to the changes in both the alienation and identification components In the case of household and equivalent incomes, however, the identification component decreased during this period www.ebook3000.com 64 On Polarization in Russia Table 5.2 The main significant changes in the DER-index (α = 0.5) over time Type of incomes Period Difference P > |t| CI 95% Individual incomes 1992/1996 1996/2005 2005/2006 2006/2007 1992/2008 1992/1996 1996/2005 1992/1996 1996/2005 0.188 –0.181 0.014 –0.030 –0.012 0.040 –0.033 0.047 –0.054 0.000 0.000 0.001 0.000 0.002 0.000 0.000 0.000 0.000 [0.175 0.202] [–0.194 –0.167] [0.006 0.023] [–0.038 –0.022] [–0.019 –0.005] [0.029 0.050] [–0.046 –0.019] [0.036 0.058] [–0.068 –0.041] Household incomes Equivalent incomes Table 5.3 The components of the DER-index (α = 0.5) Equivalent incomes Alienation Identification Correlation Individual incomes Alienation Identification Correlation 1992 1993 1994 1995 1996 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 0.415 0.446 0.481 0.492 0.540 0.491 0.468 0.445 0.435 0.426 0.423 0.409 0.432 0.399 0.404 –0.207 –0.223 –0.239 –0.232 –0.204 –0.210 –0.231 –0.221 –0.202 –0.187 –0.208 –0.201 –0.215 –0.184 –0.191 0.524 0.539 0.609 0.631 0.706 0.669 0.646 0.647 0.621 0.615 0.582 0.569 0.562 0.543 0.539 –0.182 –0.188 –0.217 –0.225 –0.261 –0.242 –0.246 –0.252 –0.222 –0.217 –0.190 –0.179 –0.176 –0.165 –0.171 identification Year 0.762 0.753 0.742 0.722 0.694 0.702 0.739 0.738 0.723 0.713 0.747 0.746 0.753 0.743 0.746 0.739 0.731 0.769 0.802 0.968 0.900 0.833 0.824 0.795 0.787 0.746 0.694 0.732 0.682 0.683 1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65 0.60 0.55 0.50 1992 1993 1994 1995 1996 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 year individual incomes total household income Fig 5.2 Identification component of the DER-index equivalent incomes 5.2 Empirical Results 5.2.2 65 Decomposition of Polarization Measures by Income Sources In the previous subsection the DER-index for the different years was computed The purpose of this subsection is to estimate the contribution of different income sources to the changes in this measure Firstly, the DER-index will be broken down using the so-called Shapley decomposition procedure Then, it will be decomposed using the approach proposed by Araar (2008) Since a decomposition by income sources analysis can include only positive incomes, we first compute the DER-index when only positive incomes are taken into account Table 5.4 and Fig 5.3 give these results Table 5.5 summarizes the main (largest) changes for the DER-index for individual (positive) incomes The significant changes over time can be divided into two sub-periods in two ways: an increase between 1992 and 1996 and a decrease between 1996 and 2005 (first decomposition in Tables 5.9 and 5.10) or an increase between 1992 and 2001 and a decrease between 2001 and 2005 (second decomposition) The shares of the various income sources in total income and the values of the Gini for each income source are given in Chap (Tables 4.8 and 4.9) Table 5.6 gives the values of the DER-index for each income source and various years Positive incomes, as expected, are less unequal and polarized The largest differences between the measures for all incomes and for positive ones were found for 1996, when the percentage of zero-incomes was the highest The lowest DER-index was found for transfers Self-employment incomes are more polarized than incomes from salaried work Other sources have the highest degree of polarization Table 5.4 The DER-index for positive incomes (α = 0.5) 1992 1993 1994 1995 1996 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 Individuals [CI 95%] Households [CI 95%] Equivalent incomes [CI 95%] 0.284 0.279 0.301 0.310 0.320 0.316 0.330 0.331 0.316 0.315 0.291 0.275 0.282 0.274 0.279 0.271 0.285 0.292 0.289 0.293 0.289 0.296 0.294 0.286 0.279 0.280 0.279 0.293 0.273 0.283 0.251 0.262 0.272 0.271 0.274 0.262 0.266 0.256 0.251 0.248 0.250 0.244 0.255 0.242 0.244 [0.279 [0.274 [0.293 [0.302 [0.311 [0.306 [0.322 [0.316 [0.308 [0.308 [0.286 [0.270 [0.277 [0.271 [0.274 0.289] 0.284] 0.309] 0.317] 0.329] 0.326] 0.339] 0.346] 0.324] 0.322] 0.297] 0.279] 0.288] 0.278] 0.284] [0.265 [0.275 [0.280 [0.278 [0.283 [0.277 [0.286 [0.282 [0.274 [0.272 [0.266 [0.269 [0.277 [0.267 [0.271 0.277] 0.296] 0.303] 0.300] 0.303] 0.300] 0.306] 0.306] 0.299] 0.286] 0.294] 0.288] 0.308] 0.279] 0.295] www.ebook3000.com [0.245 [0.252 [0.261 [0.258 [0.264 [0.251 [0.253 [0.248 [0.241 [0.241 [0.240 [0.234 [0.243 [0.235 [0.232 0.257] 0.272] 0.282] 0.283] 0.284] 0.272] 0.278] 0.265] 0.261] 0.255] 0.261] 0.253] 0.268] 0.248] 0.256] 66 On Polarization in Russia 0.40 DER-index 0.35 0.30 0.25 0.20 0.15 1992 1993 1994 1995 1996 1998 2000 2001 2002 2003 2004 2005 2006 2007 2008 year individual incomes households equivalent incomes Fig 5.3 DER-index for positive incomes Table 5.5 The main changes over time in polarization for individual positive incomes Polarization index Period Difference P > |t| CI 95% Polarization for total population (DER) 1992/1996 1996/2005 1992/2001 2001/2005 1992/2007 0.036 –0.045 0.047 –0.056 –0.010 0.000 0.000 0.000 0.000 0.004 [0.025 0.046] [–0.055 –0.035] [0.031 0.062] [–0.072 –0.040] [–0.016 –0.003] Table 5.6 The DER-index (α = 0.5) for the various income sources Income source 1992 1996 2001 2005 2007 Salaried work 0.243 [0.240 0.246] 0.280 [0.271 0.289] 0.280 [0.273 0.288] 0.260 [0.255 0.266] 0.247 [0.243 0.251] Self-employment 0.304 [0.262 0.345] 0.310 [0.295 0.325] 0.363 [0.334 0.392] 0.289 [0.272 0.305] 0.293 [0.280 0.305] State transfers 0.204 [0.200 0.208] 0.195 [0.189 0.201] 0.204 [0.198 0.209] 0.213 [0.206 0.219] 0.199 [0.194 0.203] Other incomes 0.379 [0.361 0.397] 0.386 [0.356 0.415] 0.501 [0.396 0.606] 0.456 [0.424 0.487] 0.348 [0.332 0.364] Total population 0.284 [0.279 0.289] 0.320 [0.311 0.329] 0.331 [0.316 0.346] 0.275 [0.270 0.279] 0.274 [0.271 0.278] Note Confidence intervals 95% are in the brackets 5.2 Empirical Results 67 Table 5.7 Decomposition of the DER-index (α = 0.5) by income sources using the Shapley decomposition technique Income source 1992 1996 Salaried work 0.040 0.067 (0.141) (0.209) Self-employment 0.086 0.086 (0.303) (0.269) State transfers 0.039 0.049 (0.137) (0.153) Other incomes 0.119 0.118 (0.419) (0.369) Total 0.284 (1) 0.320 (1) Note The relative contributions to the overall index 2001 2005 0.049 0.056 (0.148) (0.204) 0.100 0.069 (0.302) (0.251) 0.028 0.022 (0.085) (0.080) 0.154 0.128 (0.465) (0.465) 0.331 (1) 0.275 (1) are in the parentheses 2007 0.061 (0.223) 0.076 (0.277) 0.036 (0.131) 0.101 (0.369) 0.274 (1) Tables 5.7 and 5.8 give the contribution of various income sources to the DER index: Table 5.7 gives the results based on the Shapley decomposition procedure (see Appendix A) and Table 5.8 is derived from the approach of Araar (2008) and is based on (5.1.9) When implementing the Shapley decomposition procedure it appears that all income sources have a positive impact on the DER index, but state transfers have the smallest effect Other incomes, like in the case of bipolarization, provide the largest contribution to polarization When using Araar’s method, it appears that state transfers not make the DER greater (in some periods they reduce polarization and in others they not make any contribution) and incomes from salaried work represent the highest relative contribution Note that Shapley procedure is based on marginal changes, while that of Araar on average changes Tables 5.9 and 5.10 give the results of the decomposition of changes in the DER index as a function of changes in the income sources Both the Shapley and Araar approaches show that over the whole period the decrease in the DER was mostly due to a decrease in the absolute contribution of self-employment and other incomes, while the contribution of salaried work increased Table 5.8 Decomposition of the DER-index (α = 0.5) by income sources using the Araar (2008) method Income source Salaried work 1992 1996 0.215 0.176 (0.756) (0.548) Self-employment 0.028 0.078 (0.099) (0.243) State transfers –0.019 0.001 (–0.066) (0.003) Other incomes 0.060 0.066 (0.211) (0.206) Total population 0.284 (1) 0.320 (1) Note The relative contributions to the overall index 2001 2005 0.225 0.225 (0.680) (0.818) 0.056 0.025 (0.169) (0.091) –0.006 0.000 (–0.018) (0.000) 0.056 0.025 (0.169) (0.091) 0.331 (1) 0.275 (1) are in the parentheses www.ebook3000.com 2007 0.239 (0.869) 0.017 (0.062) –0.009 (–0.033) 0.027 (0.098) 0.274 (1) ... in the median income and in the shape of the distribution of income The purpose of Chap (“Bipolarization and the Middle Class in Russia ) is to analyze the determinants of changes in income bipolarization... 1995 and 2000 and then declined sharply between 2000 and 2005, whatever the limits of the range selected Then the size of the middle class grew significantly in 2008 and then again in 2010 While in. .. the two main factors in the creation and sustenance of a middle class are higher education and stable and secure well-paid jobs 1.2 The Russian Middle Class 1.2 The Russian Middle Class Remington