REDISTRIBUTION AND STABILIZATION OF PER CAPITA INCOME OF

REDISTRIBUTION AND STABILIZATION OF PER CAPITA INCOME OF HOUSEHOLDS BY REGIONS José Marín

THE ISSUE Primary income is redistributed through the tax and transfer system: Primary income - Income taxes + transfers = Disposable income Ø Is disposable income per capita of households by regions more equally distributed and more stable than primary income? 2

REVIEW OF THE LITERATURE • Methodological approach: data driven econometric analysis of panel data: Ø Redistributive effects, variables in levels, cross-section dimension Ø Stabilizing effects, variables in changes, time series dimension • Main result: the estimated effects are very different Ø In size, redistributive effects are higher and more precisely estimated than stabilizing effects Ø In composition, there are puzzling differences: v In the same country, some operation may contribute negatively to redistribution and positively to stabilization v The same operation may contribute in one direction to redistribution or stabilization in one country and in the contrary direction in another country • No explanations for these differences 3

MAIN FINDING OF THIS PAPER The redistributive and stabilising effects of the existing tax and transfers systems can be replicated by equivalent linear taxes Three implications • Simplify estimation method and interpretation of the results • Compare estimated with observed redistributive effects • Explain stabilisation effects as a combination of fiscal impulses (not always stabilising) and drag effects (always stabilising) 4

UNDERSTANDING THE EFFECTS OF THE LINEAR TAX T y The distribution of disposable income is a linear contraction of the distribution of primary income -I 450 line � � I y

SIMPLE ESTIMATION METHOD: LINEAR REGRESIONS Estimated redistributive effect: 24. 9% (9823/39397) = 24. 9% Income tax contribution: (2111/39397) = 5. 4 pp Current transfers contribution: (7712/39397) = 19. 6 pp 45 degrees Observed redistributive effect: 24. 2% Coefficient of variation CV(y) = 20. 7% Coefficient of variation CV(d) = 15. 7% Ratio 15. 7/20. 7 = 0. 758 Observed effect is 1 - 0. 758 = 24. 2% 6

COMPARING OBSERVED AND ESTIMATED REDISTRIBUTION INCOME INEQUALITY REDISTRIBUTION EFFECT 7

COMPARING REDISTRIBUTION EFFECTS ACROSS COUNTRIES 8

UNDERSTANDING STABILISING EFFECTS Need to separate changes in income from changes in the equivalent linear tax 9

DEFINING FISCAL IMPULSE AND DRAG EFFECTS E 2 D 2 Automatic change Observed Fiscal impulse Estimated E 1 D 1 Y 2 Automatic change: 10

EXPLAINING STABILISING EFFECTS Is the elasticity of disposable to primary income less than one? Disposable Income Growth Impulse Disposable 450 line => elasticity = 1 P&E Automatic change line C&E drag C&R Primary P&R Average Primary income growth 11

STYLISED FACTS: • Expansionary fiscal effects on average (growth of disposable income higher than primary) • Comparatively small fiscal drag (22. 7%, slope of automatic changes line 0. 773) • Countercyclical effects in all years (all points are in the countercyclical regions) • Restrictive fiscal impulses only in 2011 (only point below the automatic change line) • Fiscal impulses always stabilizing (overriding the drag only when primary growth weak) • Very systematic effects (relatively high correlation coefficient on observed effects: 0. 83) 12

INTERNATIONAL COMPARISONS OF AVERAGE EFFECTS CONCLUSIONS • Estimated redistributive effects are very close to the observed ones • • Stabilising effects are a combination of fiscal impulses and drag effects Drag effects are automatic stabilisers and very close to redistribution effects Fiscal impulses can be stabilizing or destabilizing Estimated stabilizing effects can diverge from redistributive effects due to discretionary fiscal impulses 13

RESERVE SLIDES 14

ELT RESIDUALS ARE PERSISTENT 15

CONTRIBUTIONS TO THE RESIDUALS ARE ALSO PERSISTENT 16

17

18

19

• • Expansionary fiscal effects on average (0. 5 pp) There are no observations of points in the procyclical regions (NE and SW) The fiscal drag is small, like in the US, and much smaller tan in FR and GE Stabilising effects have been determined by prudent fiscal impulses in six years (2003 -2007 and 2010) Fiscal impulses have strengthened substantially the automatic effects on average (0. 75 vs. 0. 36 slopes) Not very systematic effects (Low correlation coefficient) Very strong response to the Great Recession, with delayed effects in 2010 20

• • Neutral fiscal effects on average Procyclical impulses overriding the fiscal drag in 2001 and 2010 Second highest fiscal drag, after France Fiscal impulses less stabilising than in the UK and the US Less systematic effects (lower correlation coefficient) than in the US, but more than in the UK Very small fiscal impulses in 2006 -2008 Timid response to the Great Recession in 2009, with delayed procyclical effects in 2010 21

• • • Marginally expansionary effects on average Procyclical effects (or close) in almost every year, except 2009 Highest fiscal drag of all countries Fiscal impulses neutralising or overriding the fiscal drag in most years Very systematic effects (Highest correlation coefficient) Timid response to the Great Recession (0. 5 pp), with delayed effects in 2010 22

23

24

25

26

27

Poghosyan, T. , Senhadji, A. , and Cottarelli, C. (2015): “The role of fiscal transfers in smoothing regional shocks”, chapter 2 of Carlo Cottarelli and Martine Guerguil (Eds. ): “Designing a European Fiscal Union. Lessons from the experience of fiscal federations”, International Monetary Fund. Published by Routledge 28

29

LINEAR INCOME TAX REDISTRIBUTION 30

GINI INDEX AND ITS IMPLICIT INDIFFERENCE CURVES 31

32

33

34