Commitment to Equity (CEQ): Introduc5on, Recent Innova5ons, and the CEQ Stata Package

Transcription

1 Commitment to Equity (CEQ): Introduc5on, Recent Innova5ons, and the CEQ Stata Package Sean Higgins Co-Director of Data Center and SoDware Development, CEQ Ins5tute Post-Doctoral Fellow, UC Berkeley Poverty Global Prac5ce Summer School World Bank July 17, 2017

2 Agenda 9-9:20am Introduc1on to CEQ 9:20-10:45am What s New: Recent Innova1ons in CEQ 10:45-11am Coffee Break 11-12:30pm CEQ Stata Package 2

3 Introduc5on to CEQ

4 CEQ Ins5tute: Brief Descrip5on Mission: The CEQ Ins1tute works to reduce inequality and poverty through comprehensive and rigorous tax and benefit incidence analysis, and ac1ve engagement with the policy community Objec,ve: To measure the impact of fiscal policy on inequality and poverty across the world using a comparable framework Workstreams: Research-based policy tools Data Center Advisory and training services Bridges to policy Ø Grant from Bill & Melinda Gates Founda1on US4.9 million for 5 years ( ) 4

5 CEQ Workstreams: Tools CEQ Handbook Lus1g, Nora, editor Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty. Brookings Ins1tu1on and the CEQ Ins1tute. (Online edi1on here.) 1. Methodology 2. Implementa,on 3. Applica,ons 4. Tools CEQ Master Workbook: Excel spreadsheet to present background informa1on, assump1ons and results. CEQ Checking Protocol CEQ Stata Package 5

6 CEQ Workstreams: Data Center Data on poverty and inequality across income concepts in 30 countries can be downloaded from our Data Center Future of Data Center Expand indicators included in Data Center Expand country coverage Interac1ve graphs programmed using Tableau For countries in which it is possible: Harmonized microdata Common variable names across countries for income concepts, categories of fiscal interven1on Would allow cross-country research using rich microdata Code used to convert raw microdata to harmonized and produce CEQ Assessment Allows others to test impact of changes to assump1ons 6 Research Transparency: allows replica1on of results

7 CEQ Workstreams: Advisory and Training Events like this This mini-training offered at no cost to World Bank as part of CEQ-World Bank agreement in process of being signed 2-3 day trainings at World Bank: Feb 2015, Feb 2016, Jul 2016 Aeended by Bank staff and governments (Indonesia Ministry of Finance, South Africa Treasury) Trainings at: Ghana, Paraguay, Timor Leste Ministries of Finance Inter-American Development Bank European Commission World Bank country offices in Dominican Republic and Senegal Par1cipa1on of Ministry of Finance, Central Bank, Ministry of Development, Na1onal Sta1scs Office 7

8 CEQ Workstreams: Bridges to Policy Research collabora1ons with ADB, AfDB, CAF, IDB, IMF, ICEFI, OECD, Oxfam, UNDP, UNICEF, World Bank Agreements and partnerships with OAS, CGD Director Nora Lus1g par1cipa1on in: G20 Group on Global Financial Governance World Bank Commission on Global Poverty With IMF: Ar1cle IV and IMF program reviews Completed for Costa Rica, Guatemala, Togo, Zambia In progress for Nigeria and Swaziland Soon to begin: Benin, Tajikistan 8

9 CEQ Assessments Working on over 40 countries Covers around two thirds of the world popula1on Results available online for 30 countries in our Data Center Nonresident Research Associates and over 100 collaborators U1lized by governments Working Paper series Numerous scholarly publica1ons Journal of Development Economics World Development etc. 9

10 hmp:// 10

11 CEQ-World Bank Partnership World Bank Studies using CEQ Methodology In partnership with CEQ Institute World Bank on its own 1 Armenia 1 Albania 16 Mexico* (second round) 2 Chile 2 Armenia (second round) 17 Mongolia 3 Dominican Republic 3 Bangladesh 18 Montenegro* 4 Ethiopia 4 Belarus 19 Mozambique* 5 Georgia 5 Brazil (second round) 20 Namibia* 6 Ghana 6 Cameroon 21 Pakistan 7 Indonesia 7 Colombia 22 Poland 8 Jordan 8 Comoros* 23 Republic of Congo 9 Paraguay 9 Croatia 24 Russia (second round) 10 Russia 10 Egypt* 25 Senegal* 11 South Africa 11 Gabon 26 Serbia 12 Sri Lanka 12 Greece 27 Sri Lanka (second round) 13 Tanzania 13 Indonesia* (second round) 28 Turkey 14 Zambia 14 Latvia 29 Vietnam* 15 Mali Note: *In collaboration with CEQ Institute or with a CEQ Institute team member as consultant

12 Data, Informa5on, and SoDware Requirements Household survey (representa1ve at the na1onal level, most recent available) Input-output table or Social Accoun1ng Matrix (preferably of year close to household survey) Detailed descrip1on of each tax and spending item to be included in the analysis Budget & administra1ve data for the year of the survey Stata 13 or higher Make sure to update all To export graphs directly to Excel, Stata 14 or higher 12

13 CEQ Assessment: Income Concepts PRE-FISCAL INCOME (MARKET OR MARKET PLUS PENSIONS) PLUS DIRECT TRANSFERS MINUS DIRECT TAXES DISPOSABLE INCOME PLUS INDIRECT SUBSIDIES MINUS INDIRECT TAXES CONSUMABLE INCOME PLUS MONETIZED VALUE OF PUBLIC SERVICES: EDUCATION & HEALTH FINAL INCOME 13

14 CEQ Assessment: Fiscal Interven5ons Currently included: Direct taxes Direct cash transfers Non-cash direct transfers such as school uniforms and school lunches Contribu1ons to pensions and social insurance systems Indirect taxes on consump1on Indirect subsidies In-kind transfers such as spending on educa1on and health (valued at government cost) 14

15 Alloca5on Methods Direct Iden1fica1on from survey However, results must be checked: how realis1c are they? If informa1on not directly available in microdata, then: Inference Imputa1on Simula1on Predic1on Alternate Survey Secondary Sources (last resort) 15

16 CEQ Assessment: Ques5ons How much income redistribu1on and poverty reduc1on is being accomplished through fiscal policy? How equalizing and pro-poor are specific taxes and government spending? How effec1ve are taxes and government spending in reducing inequality and poverty? What is the impact of fiscal reforms that change the size and/or progressivity of a par1cular tax or benefit? 16

17 The image part with relationship ID rid2 was not found in the file. Composi5on of Social Spending as a Share of GDP (circa 2010) 30% (ranked by social spending plus contributory pensions / GDP; GNI right hand scale) 25,000 25% 20% 15% 10% 5% 20,000 15,000 10,000 5,000 0% OECD (2011) Average Argen1na (2012) Brazil (2009) Russia (2010) Uruguay (2009) Costa Rica (2010) South Africa (2010) Bolivia (2009) Tunisia (2010) Iran (2011) Venezuela (2013) Chile (2013) Jordan (2010) Colombia (2010) Mexico (2010) Georgia (2013) Armenia (2011) El Salvador (2011) Honduras (2011) Nicaragua (2009) Peru (2009) Ecuador (2011) Dominican Republic (2013) Ghana (2013) Tanzania (2011) Ethiopia (2011) Guatemala (2011) Sri Lanka (2010) Indonesia (2012) Uganda (2013) 0 Direct transfers Educa1on Health Other social spending Contributory pensions GNI per capita (2011 PPP) Source: Lus1g (2017) 17

18 The image part with relationship ID rid2 was not found in the file. Fiscal Policy and Inequality Contributory pensions as deferred income Market income plus pensions Disposable income Consumable income Final income Argen1na (2012) Armenia (2011) Bolivia (2009) Brazil (2009) Chile (2013) Colombia (2010) Costa Rica (2010) Dominican Republic (2013) Ecuador (2011) El Salvador (2011) Ethiopia (2011) Georgia (2013) Ghana (2013) Guatemala (2011) Honduras (2011) Indonesia (2012) Iran (2011) Jordan (2010) Mexico (2010) Nicaragua (2009) Peru (2009) Russia (2010) South Africa (2010) Sri Lanka (2010) Tanzania (2011) Tunisia (2010) Uganda (2013) Uruguay (2009) Venezuela (2013) Source: Lus1g (2017) 18

19 Fiscal Policy and Poverty Reduc5on Change in Headcount Ra5o from Market Income plus Pensions to Consumable Income (Poverty line $ PPP/day); in % Contributory pensions as deferred income (ranked by poverty reduc1on in %; poverty line $ PPP/day) 20% 0% -20% -40% 9.1% 5.9% 1.5% 1.1% 3.1% 4.1% 0.9% 1.7% 0.3% 1.1% 5.4% 0.5% 11.4% -0.7% -1.0% -0.2% -1.8% -0.8% -2.9% -2.8% -3.3% -6.5% -6.9% -4.4% -7.0% -7.3% -7.7% -8.3% -10.1% -4.7% -8.0% -10.4% -9.0% -14.9% -10.6% -15.1% -21.4% -22.0% -22.2% -25.2% -24.0% -23.3% -27.8% -28.7% -29.1% -35.0% -34.8% -35.9% -40.6% 3.3% -35.4% -13.5% -19.1% -60% -51.8% -58.4% -61.0% -51.1% -80% -71.4% -74.8% -79.4% -100% Ghana (2013) Honduras (2011) Guatemala (2011) Sri Lanka (2010) Indonesia (2012) Nicaragua (2009) Uganda (2013) Ethiopia (2011) Tanzania (2011) Dominican Republic (2013) Peru (2009) Colombia (2010) Venezuela (2013) Market income plus pensions to disposable income Armenia (2011) Tunisia (2010) Uruguay (2009) Argen1na (2012) Chile (2013) Georgia (2013) Russia (2010) Ecuador (2011) Costa Rica (2010) Jordan (2010) Brazil (2009) South Africa (2010) Mexico (2010) Bolivia (2009) El Salvador (2011) Market income plus pensions to consumable income Iran (2011) Average Source: Lus1g (2017) 19

20 What s New: Recent Innova5ons in CEQ

21 Outline of What s New in CEQ Treatment of Contributory Social Insurance Pensions Fiscal Impoverishment Indicators Effec1veness Indicators Valuing Health Benefits Valuing Educa1on Benefits Underrepor1ng and undercoverage of top incomes 21

22 Treatment of Contributory Social Insurance Pensions: Pensions as Deferred Income? (PDI) Pensions as Government Transfer? (PGT) Source: Lus1g and Higgins (2017) 22

23 Treatment of Contributory Social Insurance Pensions: Two extreme scenarios: Deferred income in actuarially fair systems: pensions included in pre-fiscal income and contribu1ons treated as mandatory savings Hence, pre-fiscal income should be net of contribu1ons Government transfer: pensions included among direct transfers and contribu1ons treated as a direct tax Hence, pre-fiscal income should be gross of contribu1ons which are subtracted out before arriving at disposable income Source: Lus1g and Higgins (2017) 23

24 Contributory Pensions: Double Coun5ng Pensions as deferred income Factor income during working years = Y Factor income during re1rement years = 0 Contribu1ons to pensions at rate s Actuarially fair system: receive pensions = sy in re1rement (for simplicity zero interest) Total direct taxes = T and benefits = B T, B in re1rement Factor income Contribu1ons Pre-fiscal Income Working age Y sy Y or (1-s)Y? Re1rement age 0 0 sy Disposable Income Source: Lus1g and Higgins (2017) 24

25 Contributory Pensions: Double Coun5ng Pensions as deferred income Factor income during working years = Y Factor income during re1rement years = 0 Contribu1ons to pensions at rate s Actuarially fair system: receive pensions = sy in re1rement (for simplicity zero interest) Total direct taxes = T and benefits = B T, B in re1rement Factor income Contribu1ons Pre-fiscal Income Disposable Income Working age Y sy (1-s)Y (1-s)Y T + B Re1rement age 0 0 sy sy T + B Source: Lus1g and Higgins (2017) 25

26 Contributory Pensions: Double Coun5ng So in PDI scenario: Pre-fiscal income is market income PLUS pensions Market income PLUS pensions is net of contribu1ons Pensions as government transfer Contribu1ons not subtracted out of pre-fiscal income Subtracted when moving to disposable (like a tax) Pre-fiscal income for re1rement age is 0 For re1red, pension added when moving to disposable income Note disposable income is the same in both scenarios Factor income Contribu1ons Pre-fiscal income Disposable Income Working age Y sy (treat as tax) Y (1-s)Y T + B Re1rement age sy T + B Source: Lus1g and Higgins (2017) 26

27 Construc5ng Income Concepts Pre-fiscal income in PGT scenario Non-Taxable Income Taxable Income Direct Transfers Contributory Pensions Pre-fiscal income in PDI scenario - Direct Taxes - Indirect Subsidies In-Kind Transfers (Educa1on, Health) Gross Income Market Income - Contribu1ons to Pensions Market Income plus Pensions Disposable Income - Net Market Income - Indirect Taxes Direct Taxes Direct Transfers Consumable Income - Copayments, User Fees 27 Final Income Source: Higgins and Lus1g (2017)

28 Outline of What s New in CEQ Treatment of Contributory Social Insurance Pensions Fiscal Impoverishment Indicators Effec1veness Indicators Valuing Health Benefits Valuing Educa1on Benefits Underrepor1ng and undercoverage of top incomes 28

29 Fiscal Impoverishment The issue: Analyzing the impact on poverty and inequality indicators can be misleading Fiscal systems can show an unambiguous reduc1on in poverty and inequality, and yet a substan1al share of the poor could have been impoverished by the combined effect of taxes and transfers Source: Higgins and Lus1g (2016) 29

30 Source: Higgins and Lus1g (2016) 30

31 Fiscal Impoverishment (Market Income plus Pensions to Consumable Income) Country (survey year) Market income plus pensions Poverty headcount (%) Change in poverty headcount (p.p.) Market income plus pensions inequality (Gini) Reynolds- Smolensky Change in inequality ( Gini) Panel A: Upper-middle income countries, using a poverty line of $ PPP per day Fiscally impoverished as % of popula,on Fiscally Impoverished as % of consumable income poor Brazil (2009) Chile (2013) Ecuador (2011) Mexico (2012) Peru (2011) Russia (2010) South Africa (2010) Tunisia (2010) Source: Higgins and Lus1g (2016) 31

32 Fiscal Impoverishment (Market Income plus Pensions to Consumable Income) Country (survey year) Market income plus pensions Poverty headcount (%) Change in poverty headcount (p.p.) Market income plus pensions inequality (Gini) Reynolds- Smolensky Change in inequality ( Gini) Fiscally impoverished as % of popula,on Fiscally Impoverished as % of consumable income poor Panel B: Lower-middle income countries, using a poverty line of $ PPP per day Armenia (2011) Bolivia (2009) Dominican Republic (2013) El Salvador (2011) Guatemala (2010) Indonesia (2012) Sri Lanka (2010) Source: Higgins and Lus1g (2016) 32

33 Fiscal Impoverishment: Axioma5c Measure The % fiscally impoverished showed earlier violates certain axioms Axioms: FI Monotonicity Focus Normaliza1on Con1nuity Permutability Transla1on invariance Linear homogeneity Subgroup consistency 33

34 Fiscal Impoverishment: Axioma5c Measure 34

35 Source: Higgins and Lus1g (2016) 35

36 Fiscal Impoverishment: Axioma5c Measure 36

37 Poverty Gap Decomposi5on: Brazil 37

38 Fiscal Impoverishment: Policy Lessons In 10 of 15 countries, between one-quarter and two-thirds of the post-fisc poor lost income to the fiscal system. In five countries, between 25 and 50% are s1ll fiscally impoverished even when the mone1zed value of educa1on and health services are included as transfers Extreme care must be taken with emphasizing domes1c resource mobiliza1on to achieve SDGs Must assess the impact on fiscal impoverishment of tax and subsidy reforms Otherwise one may not realize hur1ng a substan1al number of poor Impact on the poor of increasing taxes requires the use of adequate indicators Conven1onal measures of inequality and poverty can be awfully misleading 38 Source: Higgins and Lus1g (2016)

39 Outline of What s New in CEQ Treatment of Contributory Social Insurance Pensions Fiscal Impoverishment Indicators Effec,veness Indicators Valuing Health Benefits Valuing Educa1on Benefits Underrepor1ng and undercoverage of top incomes 39

40 Effec5veness An indicator that you typically would think of: Gini/Spending Problem: Fiscal interven1ons of larger size could be ranked worse just because higher spending may result in incrementally lower declines in Gini Decreasing marginal returns to spending for non-linear measures like Gini, squared poverty gap Leads to improper ranking of fiscal interven1ons Source: Enami (2017) 40

41 Effec5veness Addi1onal problems with Gini/Spending Not unit-free which is usually desirable for indices Measured in Gini points per $ spent Our old CEQ Effec1veness Indicator Gini/Spending/GDP is unit free but s1ll has same other issues and can be below or above 1; hard to interpret Not normalized Normaliza1on axiom: should = 1 when a program reaches its maximum efficiency Source: Enami (2017) 41

42 Effec5veness Desirable proper1es: Ranks interven1ons properly Normaliza1on Be within a certain range (i.e., between 0 and 1, or between -1 and 1) Equals 1 when program reaches maximum efficiency Intui1vely appealing interpreta1on Source: Enami (2017) 42

43 Reminder: How to Calculate the Marginal Contribu5on Let s use an example: Marginal Contribu4on of Direct Taxes to the inequality of Disposable Income Market Income Direct Taxes+Direct Transfers=Disposbale Income Two important Income concepts: Before : Disposable Income without (before subtrac1ng out) Direct Taxes o Market Income + Direct Transfers, or o Disposable Income + Direct Taxes. Aver : Disposable Income Marginal Contribu1on of the Direct Taxes: MC Direct Taxes Disposable Income = Gini Disposable Income\Direct Taxes Gini Disposable Income Direct Taxes are equalizing if MC Direct Taxes Disposable Income >0 Source: Enami (2017) 43

44 CEQ Effec5veness Indicators General Indicators: 1. Impact Effec,veness 2. Spending Effec,veness Poverty-Specific Indicators: 3. Fiscal Impoverishment and Gains Effec,veness 44

45 1. Impact Effec5veness For Inequality Indices (e.g. Gini): Impact Effectiveness T (an d o r B) End income = MC T (an d o r B) End income / MC T (an d o r B) End income, where MC T (an d o r B) End income is the maximum possible MC T (an d o r B) End income If the same amount of T is taxed op1mally to reduce inequality To achieve maximum: tax richest un1l income equal to secondrichest, tax both un1l income equal to third-richest, etc. Or same amount of B is taxed op1mally to reduce inequality To achieve maximum: give to poorest un1l income equal to secondpoorest, give to both un1l income equal to third-poorest, etc. Source: Enami (2017) 45

46 1. Impact Effec5veness For Poverty Indices (e.g. Poverty headcount ra1o): Transfers: Same formula as for inequality. Taxes can only increase poverty. New defini1on: Poverty Impact Effectiveness T End income = MC T End income / MC T End income H where MC T End income H is the Marginal Contribu1on of a tax if it is redistributed in the worst possible way. Worst possible way means tax the poorest un1l income = 0, then tax second poorest un1l income = 0, etc. So it captures how badly the poverty-increasing tax does rela1ve to the amount of harm it could poten1ally do Source: Enami (2017) 46

47 1. Impact Effec5veness This Indicator is always between -1 and +1 and the higher its value, the beeer it is. Interpreta1on: Given the amount we spent (or taxed), we achieved X% of the inequality (or poverty) reduc1on that was possible Rela1ve realized inequality or poverty reduc1on of a tax, a transfer or a combina1on of taxes and transfers Example: inequality impact effec1veness of a transfer = 0.7 à the transfer has realized 70% of its poten1al to reduce inequality In the context of poverty and only for the taxes: how much of the tax s poten1al to harm the poor was realized? (More nega1ve à more poten1al for harm realized) Source: Enami (2017) 47

48 1. Impact Effec5veness (Applica5on: Iran) Fiscal Incident Direct Taxes and Contributions Impact Effectiveness with respect to: Consumable Income Disposable Income Final Income Income Tax Employee contributions to the health insurance Employer contributions to the health insurance Employee contributions to the Social Security Employer contributions to the Social Security Total Direct Taxes and Contributions Direct Transfers Targeted Subsidy Program Social Assistance Semi-cash Transfers (Food) Total Direct Transfers Indirect Taxes (Sales Taxes) Education Transfers In-kind Transfers Education User-fees Health Transfers Health User-fees - - Note: The Gini coefficient is the index used to calculate the effectiveness indicator here Source: Enami (2017)

49 2. Spending Effec5veness It is only applicable to the taxes and transfers with posi1ve Marginal Contribu1on (inequality or poverty reducing) Spending Effectiveness T (and/or B) End income = T (and/or B )/T (and/ or B) where T (and/or B ) is the minimum amount of Tax (or Benefit) that is needed to create the same MC T (an d o r B) End income This Indicator is always between 0 and +1 and the higher its value, the beeer it is. Source: Enami (2017) 49

50 2. Spending Effec5veness (Applica5on: Iran) Fiscal Incident Spending Effectiveness with respect to: Disposable Income Consumable Income Final Income Income Tax Employee contributions to the health insurance Direct Taxes and Contributions Employer contributions to the health insurance Employee contributions to the Social Security Direct Transfers Employer contributions to the Social Security Total Direct Taxes and Contributions Targeted Subsidy Program Social Assistance Semi-cash Transfers (Food) N/A N/A N/A Total Direct Transfers Indirect Taxes (Sales Taxes) - N/A N/A In-kind Transfers Education Transfers Education User-fees Health Transfers Health User-fees - Source: - Enami (2017) N/A 50

51 3. Fiscal Impoverishment and Gains Effec5veness It is only applicable to the poverty indicators. It uses two concepts introduced in Higgins and Lus1g (2016): Fiscal Impoverishment (FI): How much poor individuals are made worse off by a fiscal system. Fiscal Gains to the Poor (FGP): How much poor individuals are made beeer off by a fiscal system. Higgins, Sean, and Nora Lus1g Can a poverty-reducing and progressive tax and transfer system hurt the poor? Journal of Development Economics 122:

52 3. Fiscal Impoverishment and Gains Effec5veness For a fiscal system (composed of taxes and transfers): Effectiveness FI/FGP =[( B/T+B )( FGP_MC T and B End income /B )]+[( T/ T+B )(1 FI_MC T and B End income /T )] where: B > 0 is total transfers, T > 0 is total taxes FGP_MC T and B End income 0 is the marginal contribu1on of T and B to FGP FI_MC Tand B End income 0 is the marginal contribu1on of T and B to FI This is a weighted average of : Tax Effectiveness FI =1 FI_MC T End income /T, Transfer Effectiveness FGP = FGP_MC B End income /B Source: Enami, Higgins, and Younger (2017) 52

53 Outline of What s New in CEQ Treatment of Contributory Social Insurance Pensions Fiscal Impoverishment Indicators Effec1veness Indicators Valuing Health Benefits Valuing Educa1on Benefits Underrepor1ng and undercoverage of top incomes 53

54 Valuing Health Benefits We follow so-called expenditure incidence or the government cost-of-provision approach Per beneficiary input costs obtained from administra1ve data as the measure of average benefits As disaggregated as possible E.g. by type of care and by state This approach amounts to asking the following ques1on: How much would the income of a household have to be increased if it had to pay for the free or subsidized public service at its full cost to the government? 54

55 Valuing Health Benefits Issue: welfare impact could be very different than amount spent For example: low-cost preventa1ve care (e.g. oral rehydra1on therapy, vaccina1on) can have large health impacts Alterna1ve: Behavioral-outcome approach Accounts for behavioral change and relies on outcomes to measure welfare Ongoing work by Jeremy Barofsky Strategy: use natural experiments where public health insurance coverage was expanded to es1mate effect of different health interven1ons on mortality Then convert to $ using value of sta1s1cal life 55

56 Valuing Health Benefits Limita1ons of the behavioral-outcome approach Necessary data and natural experiments to evaluate welfare impact not available in most countries Relies on value of sta1s1cal life es1mates Unlikely that these methods will replace government costof-provision approach in CEQ methodology In ongoing work Jeremy Barofsky is comparing the results from this method to cost-of-provision Can be added as robustness checks when possible 56

57 Outline of What s New in CEQ Treatment of Contributory Social Insurance Pensions Fiscal Impoverishment Indicators Effec1veness Indicators Valuing Health Benefits Valuing Educa,on Benefits Underrepor1ng and undercoverage of top incomes 57

58 Valuing Educa5on Benefits Same as health: government cost-of-provision approach As disaggregated as possible E.g. by level of schooling and by state Issue: welfare impact could be very different than amount spent Net present value of educa1on benefits over life1me Other possibili1es to determine benefit of public educa1on Mincer regressions but many reasons this might not es1mate private rate of return (Heckman et al 2006) 58

59 Valuing Educa5on Benefits Ongoing work by Sergio Urzua Es1mate internal rates of return by modeling decision of whether to pursue next level of educa1on Drawbacks to this approach: Leads to es1mates of differences in return across levels, not an absolute level of the return Like new health methods, unlikely to replace government cost-of-provision approach in CEQ methodology Amounts can be compared to current methodology Can be added as robustness checks when possible 59

60 Outline of What s New in CEQ Treatment of Contributory Social Insurance Pensions Fiscal Impoverishment Indicators Effec1veness Indicators Valuing Health Benefits Valuing Educa1on Benefits Underrepor,ng and undercoverage of top incomes 60

61 Underrepor5ng and Undercoverage at Top Issue: mul1ple issues lead to bias in inequality es1mates 1. Underrepor1ng of incomes Can happen anywhere in the distribu1on Don t know direc1on of bias on inequality es1mate 2. Unit non-response Rich are less likely to respond to survey Counter-intui1ve: not necessarily true that this à inequality is underes1mated (Deaton, 2005) A missing rich person, once added back into survey, affects both rela1ve distribu1on and mean income Gini is func1on of both In prac1ce, this problem has led to underes1ma1on of inequality (e.g. Hlasny and Verme, forthcoming) 61

62 Underrepor5ng and Undercoverage at Top Issue: mul1ple issues lead to bias in inequality es1mates 3. Extreme observa1ons Even in absence of underrepor1ng or higher probability of unit non-response from rich Incomes of the rich are sparse (long tail of distribu1on) Suppose our survey samples 1% of popula1on, evenly distributed throughout distribu1on We will sample 1 of richest 100 people Assuming Pen parade is convex at upper tail of distribu1on: In expecta1on, we get the right income for richest 100 More likely to underes1mate than overes1mate But if we overes1mate, expected to do so by more than if we underes1mate 62

63 Underrepor5ng and Undercoverage at Top Issue: mul1ple issues lead to bias in inequality es1mates 3. Extreme observa1ons 63

64 Underrepor5ng and Undercoverage at Top Poten1al solu1ons Reweight or adjust incomes Parametric correc1on to top incomes (e.g. fit a Pareto to upper tail of distribu1on) Use tax record tabula1on; cell-based imputa1ons Drawbacks of these: Based on assump1ons we haven t had the data to test We don t know which of the three issues described before is more prevalent Ongoing work by Facundo Alvaredo, Mauricio De Rosa, Sean Higgins, Nora Lus1g, Andrea Vigorito Merge individual-level survey and tax return data to quan1fy extent of each issue, test assump1ons and solu1ons 64

65 CEQ Stata Package

66 Outline of CEQ Stata Package Genng started and resources Treatment of Contributory Social Insurance Pensions with CEQ Stata commands Newest commands Commands to run first; check basic results Commands for more advanced tasks Ongoing work: standalone commands for CEQ indicators (fiscal impoverishment, effec1veness) 66

67 CEQ Stata Package: Gegng started Make sure have Stata 13 or newer To export graphs (ceqgraph commands) directly to MWB, need Stata 14 or newer To install or update the CEQ Stata Package: update all ssc install ceq, replace Include the above in your do files that use CEQ Stata commands This ensures always using most recent version of commands Read the resources (next slide) 67

68 CEQ Stata Package: Resources CEQ Handbook Chapter 7 (Higgins, 2017) All the indicators used in the results in MWB Sec1ons D and E Commands and their syntax If analysis separated by group: Chapter 8 (Aranda and Ratzlaff, 2017) help ceq and help files for other commands If you get an error or have sugges1ons to improve the package me at sean.higgins@ceqins1tute.org Always working on improving package For example, thanks to Mata code for Ginis and concentra1on coefficients from Paul Corral, improved efficiency and run1mes of commands 68

69 Outline of CEQ Stata Package Gezng started and resources Treatment of Contributory Social Insurance Pensions with CEQ Stata commands Newest commands Commands to run first; check basic results More commands Ongoing work: standalone commands for CEQ indicators (fiscal impoverishment, effec1veness) 69

70 Construc5ng Income Concepts Pre-fiscal income in PGT scenario Non-Taxable Income Taxable Income Direct Transfers Contributory Pensions Pre-fiscal income in PDI scenario taxable() pensions() dtransfers() education() health() otherpublic() - subsidies() gross() Direct Taxes dtaxes() - Indirect Subsidies In-Kind Transfers (Educa1on, Health) Gross Income Market Income - Contribu1ons contribs() to Pensions Market Income plus mpluspensions() Pensions Disposable Income Consumable Income - Copayments, User Fees Final Income final() - Net Market Income market() - Indirect Taxes Direct Taxes Direct Transfers disposable() consumable() dtaxes() netmarket() dtransfers() indtaxes() userfeeseduc() userfeeshealth() userfeesother() 70 Source: Higgins (2017)

71 Outline of CEQ Stata Package Gezng started and resources Treatment of Contributory Social Insurance Pensions with CEQ Stata commands Newest commands Commands to run first; check basic results More commands Ongoing work: standalone commands for CEQ indicators (fiscal impoverishment, effec1veness) 71

72 Newest Commands ceqmarg calculates the marginal contribu1on of each fiscal interven1on to inequality, poverty, reranking ceqef calculates effec1veness indicators for broad categories (going from one core income concept to another) ceqefext calculates effec1veness indicators for each fiscal interven1on ceqcoverage calculates coverage and leakages among each income group for each fiscal interven1on ceqtarget: same but among target popula1on 72

73 Outline of CEQ Stata Package Gezng started and resources Treatment of Contributory Social Insurance Pensions with CEQ Stata commands Newest commands Commands to run first; check basic results More commands Ongoing work: standalone commands for CEQ indicators (fiscal impoverishment, effec1veness) 73

74 Commands to run first; check basic results ceqppp as input to other commands Automates PPP conversions ceqassump gives inequality, poverty, distribu1on by decile Unlike other commands, no specific op1ons for each income concept Instead list any set of income variables in varlist Many uses First glance at results Test effect of different assump1ons when construc1ng income concepts Policy simula1ons 74

75 Commands to run first; check basic results ceqdes gives non-distribu1onal summary sta1s1cs For both income concepts and fiscal interven1ons % with non-0; mean; median; etc. Tip: rather than construct all income concepts first, oven teams will start using ceqassump and ceqdes as they go E.g. construct market income plus pensions and disposable income, use ceqassump and ceqdes, check these results to see if reasonable Oven even produce more results (ceqlorenz, ceqfiscal, ceqextend ) and send through CEQ Checking Protocol before construc1ng consumable, final income 75

76 Outline of CEQ Stata Package Gezng started and resources Treatment of Contributory Social Insurance Pensions with CEQ Stata commands Newest commands Commands to run first; check basic results More commands Ongoing work: standalone commands for CEQ indicators (fiscal impoverishment, effec1veness) 76

77 More commands ceqlorenz, ceqfiscal, ceqextend jointly produce a lot of the main incidence and concentra1on results for Sec1on D ceqfi produces the fiscal impoverishment indicators ceqstatsig assesses sta1s1cal significance of differences in inequality and poverty across core income concepts ceqextsig does the same for impact of par1cular fiscal interven1ons on inequality and poverty ceqgraph (with various subcommands) produces graphs of Lorenz curves, concentra1on curves, CDFs, fiscal impoverishment curves 77

78 Outline of CEQ Stata Package Gezng started and resources Treatment of Contributory Social Insurance Pensions with CEQ Stata commands Newest commands Commands to run first; check basic results More commands Ongoing work: standalone commands for CEQ indicators (fiscal impoverishment, effec,veness) 78

79 Thank you! 79

80 Credits 80

81 CEQ Ins5tute: Team TEAM Nora LusDg, Director Ludovico Feoli, Director of Policy Area Core Team (in alphabe1cal order): Maynor Cabrera, Director of Projects and Advisory Services and Associate Director for La4n America & the Caribbean Samantha Greenspun, Director of Grants and Project Management Sean Higgins, Co-Director of CEQ Data Center and SoQware Development Jon Jellema, Associate Director for Africa, Asia and Europe Carlos Mar4n-del-Campo, Director of Communica4ons Israel Mar4nez, Coordinator of CEQ Masterdata Itzel Mar4nez, Administra4ve Coordinator Sandra Mar4nez, Co-Director of CEQ Data Center and SoQware Development Estuardo Moran, Associate Director for La4n America & the Caribbean Lisa Paterson, Assistant Director Stephen Younger, Associate Director for Africa, Asia and Europe Research Associates (resident): Jim Alm, Rodrigo Aranda, Stefano Barbieri, Koray Caglayan, Enrique de la Rosa, Ali Enami, Siyu Quan Research Assistants: Marc Brooks, Cris4na Carrera, Ruoxi Li, Michael Ossorguine, Xavi Recchi 81

82 References: Aranda, Rodrigo and Adam Ratzlaff Analyzing the Impact of Fiscal Policy on Ethno-Racial Inequality. Chapter 8 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Deaton, Angus Measuring Poverty in a Growing World (or Measuring Growth in a Poor World). Review of Economics and Sta4s4cs 87: Enami, Ali Measuring the Effec1veness of Taxes and Transfers in Figh1ng Poverty and Reducing Inequality in Iran. Chapter 16 in Lus1g, Nora, editor Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty. Brookings Ins1tu1on and the CEQ Ins1tute. Enami, Ali, Sean Higgins and Stephen Younger Fiscal Impoverishment and Gains Effec1veness Indicators. Authored box in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Enami, Ali, Nora Lus1g and Rodrigo Aranda Analy1cal Founda1ons: Measuring the Redistribu1ve Impact of Taxes and Transfers, Chapter 2 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Heckman, James, Lance J. Lochner, and Petra E Todd. Earnings Func1ons, Rates of Return and Treatment Effects: The Mincer Equa1on and Beyond. In Handbook of the Economics of Educa1on. Higgins, Sean Producing Indicators and Results, and Comple1ng Sec1ons D and E of CEQ Master Workbook Using the CEQ Stata Package. Chapter 7 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Higgins, Sean and Nora Lus1g Can a Poverty-Reducing and Progressive Tax and Transfer System Hurt the Poor? Journal of Development Economics 122: Higgins, Sean and Nora Lus1g Alloca1ng Taxes and Transfers, Construc1ng Income Concepts, and Comple1ng Sec1ons A, B, and C of CEQ Master Workbook, Chapter 5 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Hlasny, Vladimir and Paolo Verme. Forthcoming. Top Incomes and the Measuremen of Inequality in Egypt. World Bank Economic Review. Lus1g, Nora

83 Teams and References by Country (the year for which the analysis was conducted in parentheses); C=consump5on & I=income) 1. Argen,na ( ; I): Rossignolo, Dario. 2017a. Taxes, Expenditures, Poverty, and Income Distribu1on in Argen1na, Chapter 10 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Rossignolo, Dario. 2017b. CEQ Master Workbook: Argen1na. Version: May 19, 2017, CEQ Data Center (CEQ Ins1tute, Tulane University). 2. Armenia (2011; I): Younger, Stephen D. and Artsvi Khachatryan Fiscal Incidence in Armenia, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g. (World Bank). Younger, Stephen D. and Artsvi Khachatryan CEQ Master Workbook: Armenia. Version: May 31, 2014, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 3. Bolivia (2009; I): Paz Arauco, Veronica, George Gray-Molina, Wilson Jimenez and Ernesto Yañez. 2014a. Explaining Low Redistribu1ve Impact in Bolivia, in The Redistribu4ve Impact of Taxes and Social Spending in La4n America, edited by Nora Lus1g, Carola Pessino and John Scoe, Special Issue, Public Finance Review 42, no 3, pp DOI: / Paz Arauco, Veronica, George Gray-Molina, Wilson Jimenez and Ernesto Yañez. 2014b. CEQ Master Workbook: Bolivia. Version: September 22, 2014, CEQ Data Center (CEQ Ins1tute, Tulane University). 4. Brazil ( ; I): H i g g i n s, S e a n a n d C l a u d i n e y P e r e i r a The Effects of Brazil s Taxa1on and Social Spending on the Distribu1on of Household Income, in The Redistribu4ve Impact of Taxes and Social Spending in La4n America, edited by Nora Lus1g, Carola Pessino and John Scoe, Special Issue, Public Finance Review 42, 3, pp DOI: / Higgins, Sean and Claudiney Pereira CEQ Master Workbook: Brazil. Version: April 19, 2017, CEQ Data Center (CEQ Ins1tute, Tulane University). 5. Chile (2013, I): Mar1nez- Aguilar, Sandra, Alan Fuchs, Eduardo Or1z-Juarez and Giselle del Carmen The Impact of Fiscal Policy on Inequality and Poverty in Chile, Chapter 12 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Mar1nez-Aguilar, Sandra and Eduardo Or1z-Juarez CEQ Master Workbook: Chile. Version: October 7, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 6. Colombia (2010, I): Melendez, Marcela and Valen1na Mar1nez CEQ Master Workbook: Colombia. Version: December 83 17, 2015, CEQ Data Center (CEQ Ins1tute, Tulane University and Inter-American Development Bank).

84 Teams and References by Country (the year for which the analysis was conducted in parentheses); C=consump5on & I=income) 7. Costa Rica (2010; I): S a u m a, P a b l o a n d J u a n D i e g o T r e j o s a. Gasto publico social, impuestos, redistribucion del ingreso y pobreza en Costa Rica, CEQ Working Paper 18 (Center for Inter- American Policy and Research and Department of Economics, Tulane University and Inter-American Dialogue), January. Sauma, Pablo and Juan D. Trejos. 2014b. CEQ Master Workbook: Costa Rica. Version: February 2014, CEQ Data Center (CEQ Ins1tute, Tulane University). 8. Dominican Republic ( , I): Aristy-Escuder, Jaime, Maynor Cabrera, Blanca Moreno-Dodson and Miguel E. Sanchez-Mar1n Fiscal Policy and Redistribu1on in the Dominican Republic, Chapter 13 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Aristy-Escuder, Jaime, Maynor Cabrera, Blanca Moreno-Dodson and Miguel Sanchez-Mar1n CEQ Master Workbook: Dominican Republic. Version: August 4, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 9. Ecuador ( , I): Llerena Pinto, Freddy Paul, Maria Cristhina Llerena Pinto, Roberto Carlos Saa Daza and Maria Andrea Llerena Pinto Social Spending, Taxes and Income Redistribu1on in Ecuador, CEQ Working Paper 28 (Center for Inter- American Policy and Research and Department of Economics, Tulane University and Inter-American Dialogue), February. Llerena Pinto, Freddy Paul, Maria Cristhina Llerena Pinto, Roberto Carlos Saa Daza and Maria Andrea Llerena Pinto CEQ Master Workbook: Ecuador. Version: January 5, 2017, CEQ Data Center (CEQ Ins1tute, Tulane University). 10. El Salvador (2011; I): Beneke, Margarita, Nora Lus1g and Jose Andres Oliva The Impact of Taxes and Social Spending on Inequality and Poverty in El Salvador, Chapter 14 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Beneke, Margarita, Nora Lus1g and Jose Andres Oliva CEQ Master Workbook: El Salvador. Version: June 26, 2014, CEQ Data Center (CEQ Ins1tute, Tulane University and Inter-American Development Bank). 11. Ethiopia (2011; C): Hill, Ruth, Gabriela Inchauste, Nora Lus1g, Eyasu Tsehaye and Tassew Woldehanna A Fiscal Incidence Analysis for Ethiopia, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g (World Bank). Hill, Ruth, Eyasu Tsehaye and Tassew Woldehanna CEQ Master Workbook: Ethiopia. Version: September 28, 2014, 84 CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank).

85 Teams and References by Country (the year for which the analysis was conducted in parentheses); C=consump5on & I=income) 12. European Union (2011, I): EUROMOD sta1s1cs on Distribu1on and Decomposi1on of Disposable Income, accessed at hep:// using EUROMOD version no. G Georgia (2013; I): Cancho, Cesar and Elena Bondarenko The Distribu1onal Impact of Fiscal Policy in Georgia," in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g (World Bank). Cancho, Cesar and Elena Bondarenko CEQ Master Workbook: Georgia. Version: December 31, 2015, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 14. Ghana ( ; C): Younger, Stephen, Eric Osei-Assibey and Felix Oppong Fiscal Incidence in Ghana. Review of Development Economics. Published electronically January 11, DOI: /rode Younger, Stephen, Eric Osei-Assibey, and Felix Oppong CEQ Master Workbook: Ghana, Version: February 10, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University). 15. Guatemala (2011; I): Cabrera, Maynor, Nora Lus1g and Hilcias E. Moran Fiscal Policy, Inequality and the Ethnic Divide in Guatemala. World Development 76 (December), pp DOI: /j.worlddev Cabrera, Maynor and Hilcias E. Moran. 2015a. CEQ Master Workbook: Guatemala. Version: May 6, 2015, CEQ Data Center (CEQ Ins1tute, Tulane University, Ins1tuto Centroamericano de Estudios Fiscales (ICEFI) and Interna1onal Fund for Agricultural Development (IFAD)). 16. Honduras (2011; I): Icefi Incidencia de la poli1ca fiscal en el ambito rural de Centro America: el caso de Honduras, CEQ Working Paper 51 (CEQ Ins1tute, Tulane University, IFAD and Ins1tuto Centroamericano de Estudios Fiscales). Castaneda, Ricardo and Ilya Espino CEQ Master Workbook: Honduras. Version: August 18, 2015, CEQ Data Center (CEQ Ins1tute, Tulane University, Ins1tuto Centroamericano de Estudios Fiscales and Interna1onal Fund for Agricultural Development). 17. Indonesia (2012; C): Jellema, Jon, Maehew Wai-Poi, and Rythia A}ar The Distribu1onal Impact of Fiscal Policy in Indonesia, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g (World Bank). A}ar, Rythia, Jon Jellema, and Maehew Wai-Poi CEQ Master Workbook: Indonesia. Version: February 26, 2015, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 85

86 Teams and References by Country (the year for which the analysis was conducted in parentheses); C=consump5on & I=income) 1 8. I r a n ( ; C & I ) : E n a m i, A l i a. Measuring the Effec1veness of Taxes and Transfers in Figh1ng Poverty and Reducing Inequality in Iran. Chapter 16 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Enami, Ali, Nora Lus1g and Alireza Taqdiri. 2017b. CEQ Master Workbook: Iran. Version: May 5, 2017, CEQ Data Center (CEQ Ins1tute, Tulane University and Economic Research Forum). 19. Jordan ( ; C): Alam, Shamma A., Gabriela Inchauste and Umar Serajuddin The Distribu1onal Impact of Fiscal Policy in Jordan, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g (World Bank). Abdel-Halim, Morad, Shamma A. Alam, Yusuf Mansur, Umar Serajuddin and Paolo Verme CEQ Master Workbook: Jordan. Version: March 8, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 20. Mexico (2010; C & I): Scoe, John Redistribu1ve Impact and Efficiency of Mexico s Fiscal System, in The Redistribu4ve Impact of Taxes and Social Spending in La4n America, edited by Nora Lus1g, Carola Pessino, John Scoe, Special Issue, Public Finance Review 42, no. 3, pp DOI: / Scoe, John CEQ Master Workbook: Mexico. Version: September 2, 2013, CEQ Data Center (CEQ Ins1tute, Tulane University). 21. Nicaragua (2009; I): Icefi Incidencia de la poli1ca fiscal en la desigualdad y la pobreza en Nicaragua, CEQ Working Paper 52 (CEQ Ins1tute, Tulane University, IFAD and Ins1tuto Centroamericano de Estudios Fiscales). Cabrera, Maynor and Hilcias E. Moran. 2015b. CEQ Master Workbook: Nicaragua. Version: October 14, 2015 CEQ Data Center on Fiscal Redistribu1on (CEQ Ins1tute, Tulane University, Ins1tuto Centroamericano de Estudios Fiscales (Icefi) and Interna1onal Fund for Agricultural Development (IFAD)). 22. Peru (2009; I): Jaramillo, Miguel The Incidence of Social Spending and Taxes in Peru. in The Redistribu4ve Impact of Taxes and Social Spending in La4n America, edited by Nora Lus1g, Carola Pessino and John Scoe, Special Issue, Public Finance Review 42, no. 3, pp DOI: / Jaramillo, M CEQ Master Workbook: Peru, August 7. CEQ Ins1tute, Tulane University. 86

87 Teams and References by Country (the year for which the analysis was conducted in parentheses); C=consump5on & I=income) 23. Russia (2010; I): Lopez-Calva, Luis Felipe, Nora Lus1g, Mikhail Matytsin and Daria Popova Who Benefits from Fiscal Redistribu1on in Russia?, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g. (World Bank). Malytsin, Mikhail and Daria Popova CEQ Master Workbook: Russia. Version: March 17, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 24. South Africa ( ; I): Inchauste, Gabriela, Nora Lus1g, Mashekwa Maboshe, Catriona Purfield, Ingrid Woolard and Precious Zikhali The Distribu1onal Impact of Fiscal Policy in South Africa, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g. (World Bank). Inchauste, Gabriela, Nora Lus1g, Mashekwa Maboshe, Catriona Purfield, Ingrid Woolard and Precious Zikhali CEQ Master Workbook: South Africa. Version: March 6, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 25. Sri Lanka (2010; C): Aruna1lake, Nisha, Gabriela Inchauste and Nora Lus1g The Incidence of Taxes and Spending in Sri Lanka, in The Distribu4onal Impact of Fiscal Policy: Experience from Developing Countries, edited by Gabriela Inchauste and Nora Lus1g (World Bank). Aruna1lake, Nisha, Camilo Gomez, Nipuni Perera and Kaushalya Aeygalle CEQ Master Workbook: Sri Lanka. Version: March 10, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University and the World Bank). 26. Tanzania ( ; C): Younger, Stephen, Flora Myamba and Kenneth Mdadila Fiscal Incidence in Tanzania. African Development Review 28, no. 3, pp DOI: / Also in CEQ Working Paper 36 (CEQ Ins1tute, Tulane University, and Ithaca College and REPOA), January. Younger, Stephen, Flora Myamba, and Kenneth Mdadila CEQ Master Workbook: Tanzania. Version: June 1, 2016, CEQ Data Center (CEQ Ins1tute, Tulane University). 27. Tunisia (2010, C): Jouini, Nizar, Nora Lus1g, Ahmed Moummi, and Abebe Shimeles Fiscal Incidence and Poverty Reduc1on: Evidence from Tunisia, Chapter 17 in Commitment to Equity Handbook. Es4ma4ng the Impact of Fiscal Policy on Inequality and Poverty, edited by Nora Lus1g (Brookings Ins1tu1on Press and CEQ Ins1tute, Tulane University). Jouini, Nizar, Nora Lus1g, Ahmed Moummi, and Abebe Shimeles CEQ Master Workbook: Tunisia. Version: October 1, 2015, CEQ Data Center on Fiscal Redistribu1on (CEQ Ins1tute, Tulane University and African Development Bank). 87