The Weak-Form Efficiency Of Metal Markets Using Stationarity Tests

Kara A., Yıldırım Kasap D., Tunç G. İ.

ICOAEF’19, V. International Conference on Applied Economics and Finance , Girne, Cyprus (Kktc), 8 - 10 April 2019, pp.43

  • Publication Type: Conference Paper / Summary Text
  • City: Girne
  • Country: Cyprus (Kktc)
  • Page Numbers: pp.43
  • Middle East Technical University Affiliated: Yes


In the paper, quarterly real global prices of aluminum, copper, lead, tin, nickel, zinc, gold, silver and platinum are analyzed by KPSS-type stationarity tests considering sharp (Carrion-i-Silvestre and Sanso, 2007) and smooth breaks (Becker et al., 2006) in order to comment about the weak-form efficiency of selected metal markets. Becker et al. (2006) proposed a stationarity test which did not require any presumptions related to number, type, magnitude and form of structural breaks. This property was provided by adding best fitting fourier series to the regression equation. Since Becker et al. (2006) test has better performance for smooth breaks compared to sharp breaks, the methodology of Carrion-i-Silvestre and Sanso (2007) which allows two endogenously determined sharp breaks is also employed. Breaks may affect a macroeconomic variable either abruptly or in a slowly culminating manner. In this sense, a sharp break example would be the declaration of International Tin Council on October 24, 1985 that its funds to support the tin price exhausted. This event immediately disrupted the world tin market. On the other hand, an example for a smooth break would be the introduction of Exchange-Traded Funds backed by physical gold on March 28, 2003. This event positively affected the gold price as exchangetraded funds backed by gold became more common. The rationale of allowing breaks in the stationarity analyis is to hinder false rejections if the null hypothesis of stationarity is true. Furthermore, Sul et al. (2005)’s prewhitening approach, which limits the estimated long run variance (LRV) used in the calculation of test statistics, in order to overcome the sensitivity of KPSS type tests to LRV estimation. The standard ADF and KPSS test supported the stationarity for zinc and aluminum only. The stationarity null hypothesis of the tests proposed by Becker et al. (2006) and Carrion-i-Silvestre and Sanso (2007) was not rejected for most of the metal markets considered in the paper. This result pointed out the inefficiency of most of the metal markets and highlighted the importance of considering breaks in metal markets, especially for longer periods of time.