International stock market efficiency: a non-Bayesian time-varying model approach

Abstract

This article develops a non-Bayesian methodology to analyse the time-varying structure of international linkages and market efficiency in G7 countries. We consider a non-Bayesian time-varying vector autoregressive (TV-VAR) model, and apply it to estimate the joint degree of market efficiency in the sense of Fama (1970, 1991). Our empirical results provide a new perspective that the international linkages and market efficiency change over time and that their behaviours correspond well to historical events of the international financial system.

Keywords:

• non-Bayesian time-varying VAR model
• international linkages
• degree of market efficiency
• efficient market hypothesis

JEL Classification:

• C32
• G14
• G15

Acknowledgements

We would like to thank the editor, Mark Taylor, two anonymous referees, Kenjiro Hirayama, Jun Ma, Daisuke Nagakura, Toshiaki Watanabe, Wako Watanabe, Tomoyoshi Yabu, Yohei Yamamoto, Makoto Yano, seminar participants at Chukyo University Institute of Economics, Doshisha University, conference participants at the Japanese Economics Association 2012 Autumn Meeting in Kyushu Sangyo University, the 10th Biennial Pacific Rim Conference of the Western Economic Association International in Tokyo, the 88th Annual Conference of the Western Economic Association International in Seattle, and the Japanese Economics Association 2013 Autumn Meeting in Kanagawa University for their helpful comments and suggestions. All data and programs used for this article are available upon request.

Funding

This work was supported by the Japan Society for the Promotion of Science Grant in Aid for Scientific Research [grant number 24530364] and the Zengin Foundation for Studies on Economics and Finance [grant number 1321].

Supplemental data

Supplemental data for this article can be accessed http://dx.doi.org/10.1080/00036846.2014.909579.

Notes

¹ As described in Section II, the idea of joint efficiency is not new; for example, MacDonald and Taylor (1989) focus on the relationship between cointegration and joint efficiency in foreign exchange markets.

² Supplemental data is available at http://at-oda.com/appendix/inter_market_appendix.pdf

³ This view, however, is questioned by Dwyer and Wallace (1992).

⁴ See Supplemental material (A.1).

⁵ See Supplemental material (A.3) for more details.

⁶ In case of univariate data, this norm is essentially same as the degree defined by Ito et al. (2012). That is, the degree in this article is a natural extension of the one in Ito et al. (2012).

⁷ Table A.1 in Supplemental material provides the results of a univariate AR process for each country. From ‘Lc’ statistics, we can confirm that the TV model is more appropriate for all countries.

⁸ While we use the bootstrap method (i.e., using estimated residuals) to compute the confidence bands, the Monte Carlo method (i.e., using random draws from the normal distribution) can generate pretty much the same confidence bands. Therefore, our empirical results presented in this article do not depend on which method is used to compute the statistical inference for the degree of market efficiency.

⁹ Ito et al. (2012) utilize a much longer sample period for the US, and find that the world financial crisis in 2008 did not cause an inefficient market in the US. According to their study, inefficient markets emerged during: (i) the longest recession defined by NBER (1873–1879); (ii) the 1902–1904 recession; (iii) the New Deal era and (iv) just after the very severe 1957–1958 recession.

¹⁰ In such a case, the efficiency condition does not hold even with people’s rational behaviour.