Skip to main content

Advertisement

Log in

Revisiting comparisons of income inequality when Lorenz curves intersect

  • Original Paper
  • Published:
Social Choice and Welfare Aims and scope Submit manuscript

Abstract

The main contribution of Davies and Hoy (Am Econ Rev 85:980–986, 1995) was a “necessary and sufficient” condition for comparing inequality between income distributions according to the principle of transfer sensitivity (PTS). Chiu (Soc Choice Welf 28:375–399, 2007) showed that although the condition is sufficient, it is not necessary. In this paper, we provide the correct necessary and sufficient condition, and demonstrate with a simple example how the corrected condition allows for more pairs of distributions to be ranked by PTS. The correction clarifies the connection between Lorenz curve comparisons and inequality rankings when the curves intersect.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Similar content being viewed by others

Notes

  1. In the favourable composite transfer defined by Shorrocks and Foster (1987), the progressive transfer occurs in a lower range of incomes than does the regressive transfer, although the two transfers may occur over intersecting income ranges. See Davies and Hoy (1995) for a discussion of the normative appeal of placing greater emphasis on inequality that occurs for relatively lower incomes within the distribution.

  2. In the literature, Proposition 2 of Davies and Hoy (1995) is usually cited as an equivalence between PTS and the comparison of variances for appropriately truncated subpopulations of a pair of income distributions. See, for example, Davies and Shorrocks (2000) and Cowell (2011) for citations in monographs, and Aaberge (2009), Gajdos and Weymark (2012) and Ibragimov et al. (2018) for citations in other publications.

  3. This is clear upon comparing his Lemma 2 and Footnote 11 with our Proposition 1 here.

  4. Theorem 5.2.3 in Athreya and Lahiri (2006, p.155) states integration by parts under mild conditions. Since cumulative distribution functions are always nondecreasing right-continuous and the quadratic function in our derivations never has any point of discontinuity, it follows as a corollary of Theorem 5.2.3 that in our situation, integration by parts is workable for general distributions including discontinuities.

  5. It can be shown that \(z_i^*\) is also the unique zero root of S(y) on \([G^{-1}(P_i),F^{-1}(P_i)]\). Indeed, applying integration by parts to (2) yields \(P_iG^{-1}(P_i)+\int _{(G^{-1}(P_i),z_i^*]}G(y)dy=P_iF^{-1}(P_i)-\int _{(z_i^*,F^{-1}(P_i)]}F(y)dy\), and to \(\int _{[\underline{y},G^{-1}(P_i)]}ydG(y)=\int _{[\underline{y},F^{-1}(P_i)]}ydF(y)\) yields \(\int _{[\underline{y},G^{-1}(P_i)]}G(y)dy-P_iG^{-1}(P_i)=\int _{[\underline{y},F^{-1}(P_i)]}F(y)dy-P_iF^{-1}(P_i)\). Summing up the two equations results in \(S(z_i^*)=0\).

References

  • Aaberge R (2009) Ranking intersecting Lorenz curves. Soc Choice Welf 33:235–259

    Article  Google Scholar 

  • Athreya KB, Lahiri SN (2006) Measure theory and probability theory. Springer Texts in Statistics, 1st edn. Springer, New York, NY. https://doi.org/10.1007/978-0-387-35434-7

    Book  Google Scholar 

  • Atkinson AB (1970) On the measurement of inequality. J Econ Theory 2:244–263

    Article  Google Scholar 

  • Atkinson AB (2008) More on the measurement of inequality. J Econ Inequal 6:277–283

    Article  Google Scholar 

  • Blackorby C, Donaldson D (1978) Measures of relative equality and their meaning in terms of social welfare. J Econ Theory 18:59–80

    Article  Google Scholar 

  • Chiu WH (2007) Intersecting Lorenz curves, the degree of downside inequality aversion, and tax reforms. Soc Choice Welf 28:375–399

    Article  Google Scholar 

  • Cowell F (2011) Measuring inequality. Oxford University Press, Oxford

    Book  Google Scholar 

  • Davies J, Hoy M (1995) Making inequality comparisons when Lorenz curves intersect. Am Econ Rev 85:980–986

    Google Scholar 

  • Davies J, Shorrocks AF (2000) The distribution of wealth. In: Atkinson A, Bourguignon F (eds) Handbook of income distribution, vol 1. North-Holland, Amsterdam

    Google Scholar 

  • Gajdos T, Weymark JA (2012) Introduction to inequality and risk. J Econ Theory 147:1313–1330

    Article  Google Scholar 

  • Ibragimov M, Ibragimov R, Kattuman P, Ma J (2018) Income inequality and price elasticity of market demand: the case of crossing Lorenz curves. Econ Theor 65:729–750

    Article  Google Scholar 

  • Menezes C, Geiss C, Tressler J (1980) Increasing downside risk. Am Econ Rev 70:921–932

    Google Scholar 

  • Shorrocks AF, Foster JE (1987) Transfer sensitive inequality measures. Rev Econ Stud 54:485–497

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

Authors

Corresponding author

Correspondence to Lin Zhao.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Michael Hoy acknowledges financial support from Social Sciences and Humanities Research Council (SSHRC) (Grant 430407). Lin Zhao acknowledges financial support from the National Natural Science Foundation of China (Grant 71871212)

Appendix: Proof of the Claim

Appendix: Proof of the Claim

Decompose the population into \((n+1)\) disjoint subpopulations defined by \((P_{i-1},P_i]\). Denote the cumulative distribution function of income for the ith subpopulation under F by \(F_i\). Mathematically, \(F_i(y)\) equals 0 for \(y<F^{-1}(P_{i-1})\), equals \([F(y)-P_{i-1}]/[P_i-P_{i-1}]\) for \(y\in [F^{-1}(P_{i-1}),F^{-1}(P_i)]\), and equals 1 for \(y>F^{-1}(P_i)\). Then, \(F(y)=\sum _{i=1}^{n+1}(P_i-P_{i-1})F_i(y)\). Similarly, we define \(G_i\) and have \(G(y)=\sum _{i=1}^{n+1}(P_i-P_{i-1})G_i(y)\). Let \(L(F;P)=\int _{\underline{y}}^{F^{-1}(P)}ydF(y)/E(F)\) denote the Lorenz curve coordinate corresponding to the lowest 100P percent of income recipients. Similar notation applies to G. Under \(E(F)=E(G)\), we have: \(E(F_i)=E(G_i)\) for all \(i=1,\ldots ,n+1\); \(L(F_i;P)\ge L(G_i;P)\) for all \(P\in [0,1]\) for i odd; and \(L(F_i;P)\le L(G_i;P)\) for all \(P\in [0,1]\) for i even. The change from \(G_i\) to \(F_i\) consists of a series of MPCs if i is odd, and a series of MPSs if i is even.

Fig. 1
figure 1

An illustration of \(S_i(y)\) with \(i=1,2,3,4\)

Let \(S_i(y)=\int _{\underline{y}}^y[G_i(x)-F_i(x)]dx\). Due to \(dL(F;P)/dP=F^{-1}(P)/E(F)\) and \(dL(G;P)/dP=G^{-1}(P)/E(G)\), we have \(F^{-1}(P_i)\le G^{-1}(P_i)\) for i odd and \(G^{-1}(P_i)\le F^{-1}(P_i)\) for i even. Thus, \(S_i(y)\ge 0\) on \([G^{-1}(P_{i-1}),G^{-1}(P_i)]\) for i odd, and \(S_i(y)\le 0\) on \([F^{-1}(P_{i-1}),F^{-1}(P_i)]\) for i even. Figure 1 provides a graphical illustration of \(S_i(y)\) for \(n=3\). Observing \(S(y)=\int _{\underline{y}}^y[G(x)-F(x)]dx=\sum _{i=1}^{n+1}(P_i-P_{i-1})S_i(y)\), we have \(S(y)=(P_i-P_{i-1})S_i(y)\ge 0\) on \([F^{-1}(P_{i-1}),F^{-1}(P_i)]\) for i odd and \(S(y)\le 0\) on \([G^{-1}(P_{i-1}),G^{-1}(P_i)]\) for i even. Thus, to check (3) for all \(z\in [\underline{y},\overline{y}]\), we only need to check it on \([G^{-1}(P_i),F^{-1}(P_i)]\) with i even. \(\square \)

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Davies, J., Hoy, M. & Zhao, L. Revisiting comparisons of income inequality when Lorenz curves intersect. Soc Choice Welf 58, 101–109 (2022). https://doi.org/10.1007/s00355-021-01343-w

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s00355-021-01343-w

Navigation