24.02.2013 Views

Optimality

Optimality

Optimality

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

Modeling inequality, spread in multiple regression 129<br />

where δi(u) = F −1<br />

0 (1−(1−u) ∆i ). To estimate µ(u|xi), we let<br />

be the jumps of ˆ F0(·); then<br />

bi = ˆ F0(Y (i))− ˆ F0(Y (i−1)) =<br />

ˆµ(u|x i) = ˆ ∆ −1<br />

i<br />

�<br />

j<br />

i−1 �<br />

j=1<br />

i−1 �<br />

ˆαj = (1− ˆαi)<br />

j=1<br />

bjY (j)(1− ˆ F0(Y (j))) ˆ ∆ −1<br />

i −1<br />

where the sum is over j with ˆ F0(Y (j))≤1−(1−u) ˆ ∆i . Finally,<br />

and<br />

ˆL(u|x) = ˆµ(u|x)/ˆµ(1|x), ˆ B(u|x) = ˆ L(u|x)/u,<br />

Ĉ(u|x) = ˆµ(u|x)/ ˆ F −1 (u|x), ˆ D(u|x) = Ĉ(u|x)/u,<br />

where ˆ F −1 (u|x) is the estimate of the conditional quantile function obtained from<br />

(4.3) by replacing ∆i with ˆ ∆i and F0 with ˆ F0.<br />

Remark. The methods outlined here for the Cox proportional hazard model have<br />

been extended to the case of ties among the responses Y i, to censored data, and<br />

to time dependent covariates (see e.g. Cox [10], Andersen et al. [4] and Kalbfleisch<br />

and Prentice [16]). These extensions can be used in the analysis of the semiparametric<br />

generalized Pareto model with tied wages, censored wages, and dependent<br />

covariates.<br />

References<br />

[1] Aaberge, R. (1982). On the problem of measuring inequality (in Norwegian).<br />

Rapporter 82/9, Statistics Norway.<br />

[2] Aaberge, R. (2000a). Characterizations of Lorenz curves and income distributions.<br />

Social Choice and Welfare 17, 639–653.<br />

[3] Aaberge, R. (2000b). Ranking intersecting Lorenz Curves. Discussion Paper<br />

No. 412, Statistics Norway.<br />

[4] Andersen, P. K., Borgan, Ø. Gill, R. D. and Keiding, N. (1993).<br />

Statistical Models Based on Counting Processes. Springer, New York.<br />

[5] Atkinson, A. B. (1970). On the measurement of inequality, J. Econ. Theory<br />

2, 244–263.<br />

[6] Barlow, R. E. and Proschan, F. (1965). Mathematical Theory of Reliability.<br />

Wiley, New York.<br />

[7] Bickel, P. J. and Lehmann, E. L. (1976). Descriptive statistics for nonparametric<br />

models. III. Dispersion. Ann. Statist. 4, 1139–1158.<br />

[8] Bickel, P. J. and Lehmann, E. L. (1979). Descriptive measures for nonparametric<br />

models IV, Spread. In Contributions to Statistics, Hajek Memorial<br />

Volume, J. Juneckova (ed.). Reidel, London, 33–40.<br />

[9] Birnbaum, S. W., Esary, J. D. and Marshall, A. W. (1966). A stochastic<br />

characterization of wear-out for components and systems. Ann. Math. Statist.<br />

37, 816–826.<br />

[10] Cox, D. R. (1972). Regression models and life tables (with discussion). J. R.<br />

Stat. Soc. B 34, 187–220.<br />

ˆαj

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!