• Bernoulli
  • Volume 7, Number 2 (2001), 343-350.

Remarks on the maximum correlation coefficient

Amir Dembo, Abram Kagan, and Lawrence A. Shepp

Full-text: Open access


The maximum correlation coefficient between partial sums of independent and identically distributed random variables with finite second moment equals the classical (Pearson) correlation coefficient between the sums, and thus does not depend on the distribution of the random variables. This result is proved, and relations between the linearity of regression of each of two random variables on the other and the maximum correlation coefficient are discussed.

Article information

Bernoulli, Volume 7, Number 2 (2001), 343-350.

First available in Project Euclid: 25 March 2004

Permanent link to this document

Mathematical Reviews number (MathSciNet)

Zentralblatt MATH identifier

correlation linear regression maximum correlation spherically symmetric distributions sums of independent random variables


Dembo, Amir; Kagan, Abram; Shepp, Lawrence A. Remarks on the maximum correlation coefficient. Bernoulli 7 (2001), no. 2, 343--350.

Export citation


  • [1] Breiman, L. and Friedman, J. (1985) Estimating optimal transformations for multiple regression and correlation (with discussion). J. Amer. Statist. Assoc., 80, 580-619.
  • [2] Czáki, P. and Fisher, J. (1963) On the general notion of maximum correlation. Magyar Tudományos Akad. Mat. Kutató Intézetenk Közleményei (Publ. Math. Inst. Hungar. Acad. Sci.), 8, 27-51.
  • [3] Eaton, M. (1986) A characterization of spherical distributions. J. Multivariate Anal., 20, 272-276.
  • [4] Efron, B. and Stein, C. (1981) The jackknife estimate of variance. Ann. Statist., 9, 586-596.
  • [5] Fang, K.-T., Kotz, S., Ng, K.-W. (1990) Symmetric Multivariate and Related Distributions. London: Chapman & Hall.
  • [6] Gebelein, H. (1941) Das statistische Problem der Korrelation als Variations- und Eigenwertproblem und sein Zusammenhang mit der Ausgleichsrechnung. Z. Angew. Math. Mech., 21, 364-379.
  • [7] Gupta, A. and Varga, T. (1993) Elliptically Contoured Models in Statistics. Dordrecht: Kluwer.
  • [8] Kagan, A.M., Linnik, Yu.V. and Rao, C.R. (1973) Characterization Problems in Mathematical Statistics. Wiley.
  • [9] Lancaster, H.O. (1957) Some properties of the bivariate normal distribution considered in the form of a contingency table. Biometrika, 44, 289-292.
  • [10] Rényi, A. (1959) On measures of dependence. Acta Math. Acad. Sci. Hungar., 10, 441-451.
  • [11] Sarmanov, O.V. (1958a) Maximum correlation coefficient (symmetric case). Dokl. Akad. Nauk SSSR, 120, 715-718 (in Russian).
  • [12] Sarmanov, O.V. (1958b) Maximum correlation coefficient (non-symmetric case). Dokl. Akad. Nauk SSSR, 121, 52-55 (in Russian).
  • [13] Szekely, G.J. and Gupta, A.K. (1998) On a paper of V.B. Nevzorov. Math. Meth. Statist., 2, 122.
  • [14] Vershik, A.M. (1964) Some characteristic properties of Gaussian stochastic processes. Theory Probab. Appl., 9, 353-356 (in Russian).